Structure Elucidation of (+)-Amphidinolide A by Total Synthesis and NMR Chemical Shift Analysis

Article abstract of DOI:10.1021/ja049292k

The structure elucidation of (+)-amphidinolide A, a cytotoxic macrolide, has been accomplished by employing a combination of NMR chemical shift analysis and total synthesis. Using the reported structure as a starting point, a number of diastereomers of am

Full text of DOI:10.1021/ja049292k

Published on Web 03/30/2004
Structure Elucidation of (+)-Amphidinolide A by Total Synthesis and NMR
Chemical Shift Analysis
Barry M. Trost* and Paul E. Harrington
Department of Chemistry, Stanford UniVersity, Stanford, California 93405-5080
Received February 9, 2004; E-mail: bmtrost@stanford.edu
Table 1. Deviation of the 1H NMR Chemical Shifts of Isomers 1,
The cytotoxic macrolide, amphidinolide A, was isolated by
Kobayashi from the culture broth of the marine dinoflagellate
2, and 4-11 Relative to the Values Reported for the Isolated
Materiala
Amphidinium sp., which is symbiotically associated with an
Okinawan flatworm.1 The gross structure and relative stereochem-
istry were proposed after extensive 2D NMR experiments. However,
subsequent efforts at total synthesis revealed that the reported
structure was incorrect.2 As shown in Table 1, there were significant
1
deviations in the H NMR. Because the relative stereochemistry
was assigned with NOE data measured on a macrolide possessing
considerable flexibility, an error in the relative stereochemistry and
not gross structure seemed likely. Additionally, the differences in
chemical shifts and coupling constants are not as large as would
be expected for an error in connectivity. Unfortunately, only an
extremely small sample of the natural material remains; thus
additional NMR experiments are not possible. Therefore, total
synthesis represents the only practical method by which the correct
structure can be determined unambiguously. For such an approach
to be feasible, the synthetic route must be convergent, efficient,
and, ideally, have a minimal reliance on the chiral pool. The
synthesis of 1 previously reported by this group2c satisfies these
requirements (Figure 1).
Initially, it was assumed that the error in relative stereochemistry
was in the epoxide region where the acyclic nature of the side chain
would result in the least reliable NOE data. The trans stereochem-
istry of the epoxide was assumed to be correct on the basis of a
good correlation between the reported JH20/H21 value and other trans
epoxides. The possibility of an error in correlating the tetraol and
epoxide portions was also considered.3 These were some of the
a Spectra were measured in CDCl3 at 500 MHz. Differences are reported
assumptions made by Maleczka2b and Pattenden2a who prepared
in ppm. Values in black represent deviations of <0.04, blue values represent
0.04-0.10, green values represent 0.11-0.20, and red italicized values
represent >0.20.
isomers 2 and 3, respectively (Figure 2). However, neither matched
the data reported for the natural product.
Along these lines, we prepared C22 epimer 4 and the isomer 5
from inversion of the C19-C21 triad. Although neither matched,
the JH18/H19 value for 5 was 10.3 Hz, whereas the values for 1, 2,
4, and the isolated material were 3.3-3.8 Hz, suggesting the
requirement for trans (as drawn) C18-C19 stereochemistry.
Proceeding on the belief that the correlation of the tetraol to the
epoxide was tenuous, isomers 6-8 were prepared, combining
changes in the epoxide with a change in tetraol configuration.
However, none matched the reported data. The JH18/H19 value for 8
was 10.3 Hz, whereas 7 and 9 were both 3.4 Hz, further confirming
the requirement for trans C18-C19 stereochemistry.
Figure 1. Retrosynthetic analysis of amphidinolide A.
the site of the variations. Therefore, an error in the relative
stereochemistry of the tetraol appeared likely.
1
A comparison of the H NMR data for 1, 2, and 4-8 to the
Although we were concerned that errors within the epoxide and
tetraol would make the possibilities so numerous and complex that
we would be faced with nearly an impossible task, key spectral
data provided direction. Because the JH8/H9 and JH11/H12 values for
1, 2, and 5-9 were consistent with the natural product, it appeared
likely that the relative stereochemistry of the diols was correct, but
the stereochemistry of the C8-C9 diol was incorrect relative to
the C11-C12 diol. Therefore, isomer 9, with the C8-C9 diol
inverted, became the primary target.
natural product led us to a troubling conclusion. Initially, compari-
sons to the natural product focused on coupling constants, and, apart
from the requirement for trans C18-C19 stereochemistry, little
information was gleaned by this approach. However, as shown in
Table 1, a significant departure in the chemical shifts of the H9
and H11 tetraol protons was measured. Additionally, the differences
were not random in sign or magnitude. This was surprising because
relatively small differences were measured in the epoxide region,
9
5028
J. AM. CHEM. SOC. 2004, 126, 5028-5029
10.1021/ja049292k CCC: $27.50 © 2004 American Chemical Society

C O M M U N I C A T I O N S
Scheme 1. Synthesis of Amphidinolide A Isomer 11a
a (a) (i) (COCl)2, DMSO; (ii) Et3N (Moffatt-Swern); (b) n-BuLi,
HCtCCH2CH2OTBS, ClTi(Oi-Pr)3; (c) Red-Al, 51% (three steps); (d)
Dess-Martin, 86%; (e) Sharpless AD; (f) 1,1-dimethoxycyclopentane,
TsOH‚H2O; (g) Ph3PMeBr, NaHMDS, 60% (three steps); (h) TBAF, 78%;
(i) Moffatt-Swern; (j) Ph3PMeBr, n-BuLi, 79%; (k) DDQ, 98%; (l)
Moffatt-Swern; (m) (MeO)2POC(dN2)COMe, K2CO3, 87% (two steps);
(n) 17 (5 equiv), [Cp*Ru(MeCN)3]PF6, 23% (39% brsm); (o) piperidine,
88%; (p) AcOH/H2O (3:1); (q) TESOTf, i-Pr2NEt, 83% (two steps); (r) (i)
[RuCl2(p-cymene)]2, HCtCOEt; (ii) 20, CSA, 51%;7 (s) TBAF, AcOH,
79%; (t) [Cp*Ru(MeCN)3]PF6, 33% (38% brsm).
Figure 2. Amphidinolide A isomers 2-11.
However, isomer 9, with the C8-C9 diol inverted and epoxide
stereochemistry of 1, failed to match. Although the chemical shift
of H11 was closer than was the case for the previous isomers, H9
was 0.30 ppm upfield from the natural product. However, as shown
in Table 1, 10, the C18-C22 epimer of 9, provided an excellent
match in the tetraol region. Only in the epoxide region were the
shifts significantly different from the natural product.
identical in sign, but slightly higher than the reported value [R]24
D
+46° (c 1.0, CHCl3), therefore establishing the absolute stereo-
chemistry.
In conclusion, we have employed a combination of synthesis
and NMR spectroscopy as tools to determine the correct structure
of amphidinolide A. Although the lack of a sample of the natural
product prevents a definitive comparison, the excellent correlation
of 11 strongly suggests it is (+)-amphidinolide A.
Reexamination of the data in Table 1 indicated that 7 was an
excellent match in the epoxide region. Therefore, isomer 11, the
combination of the relative stereochemistry found in the epoxide
of 7 and the tetraol of 10, became a priority. The more quantitative
analysis of the data in Table 1 that follows also pointed to 11. The
relationship between 10 and 11 is analogous to that between 2 and
7, inversion of the C20-C22 triad. If the changes in chemical shift
that occur when the C20-C22 triad of 2 is inverted, thus yielding
7, are applied to 10, a nearly perfect match to the natural material
is obtained. For example, the chemical shift of H19 in 2 and 7 is
4.58 and 4.67 ppm, respectively. This represents a downfield shift
of 0.09 ppm. The shift of H19 in 10 is 4.65 ppm. A 0.09 ppm
downfield shift yields a predicted shift of 4.74 ppm for H19 of 11.
This value compares well to the shift of 4.72 ppm for H19 of the
natural product. Analysis of the other protons yields similar results.
The new tetraol required a complete redesign of nearly all stages
of our original synthesis.2c Ester 18 was prepared in 15 steps from
124 (Scheme 1). Conversion of 18 to 11 required a significant
change to the end game due to the sensitivity of the epoxide in 11
to acidic hydrolysis. After deprotection of 21, [Cp*Ru(MeCN)3]-
PF6-catalyzed macrocyclization of 22 provided 11, illustrating the
remarkable chemoselectivity of the Ru-catalyzed alkene-alkyne
addition. The spectral data for 11 provided an excellent fit to the
natural product. One proton deviated by 0.03 ppm, one by 0.02
ppm, and the remainder by 0.01 ppm or less. The 1H NMR spectra
in C6D6 and CD3OD deviated by 0.01 ppm or less from the isolated
material in those solvents.5 The 13C NMR spectrum deviated by
0.1 ppm or less in CDCl3.6 The J values in all three solvents were
also in agreement. These results are well within experimental error.
Acknowledgment. We thank the National Science Foundation
and the National Institutes of Health (GM-33049) for their generous
support of our program. Mass spectra were provided by the Mass
Spectrometry Facility, University of San Francisco, supported by
the NIH Division of Research Resources. We thank Professor
Robert Maleczka and Dr. Lamont Terrell for sharing their spectral
data for isomer 2.
Supporting Information Available: Experimental procedures for
11, 13-16, and 18-21. Characterization data for 4-11, 13-16, and
18-21 and spectra for 11 (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
References
(1) (a) Kobayashi, J.; Ishibashi, M.; Nakamura, H.; Ohizumi, Y.; Yamasu,
T.; Sasaki, T.; Hirata, Y. Tetrahedron Lett. 1986, 27, 5755. (b) Kobayashi,
J.; Ishibashi, M.; Hirota, H. J. Nat. Prod. 1991, 54, 1435.
(2) (a) Lam, H. W.; Pattenden, G. Angew. Chem., Int. Ed. 2002, 41, 508. (b)
Maleczka, R. E.; Terrell, L. R.; Geng, F.; Ward, J. S., III. Org. Lett. 2002,
4, 2841. (c) Trost, B. M.; Chisholm, J. D.; Wrobleski, S. A.; Jung, M. J.
Am. Chem. Soc. 2002, 124, 12420 and references therein.
(3) The epoxide to tetraol correlation was determined sequentially by NOE
between H11 and H13, H13 and H15, and H15 and H18.
(4) Alcohol 12 was prepared in three steps from (-)-diethyl tartrate: Rzepecki,
P. W.; Prestwich, G. D. J. Org. Chem. 2002, 67, 5454.
(5) In C6D6, protons 6a, 6b, 8, 9, 11, 12, 17b, 27a, and 28a deviated by 0.01
ppm. In CD3OD, protons 4, 6a, 6b, 12, and 28b deviated by 0.01 ppm.
(6) Carbons 3-5, 9, 13, 22, 24, 25, and 27-29 deviated by 0.1 ppm.
(7) Kita, Y.; Maeda, H.; Omori, K.; Okuno, T.; Tamura, Y. Synlett 1993,
273.
Finally, the optical rotation [R]24 +56° (c 0.05, CHCl3) was
JA049292K
D
9
J. AM. CHEM. SOC. VOL. 126, NO. 16, 2004 5029