Enantioselective Total Synthesis of Mandelalide A and Isomandelalide A: Discovery of a Cytotoxic Ring-Expanded Isomer

Article abstract of DOI:10.1021/jacs.5b12318

The total synthesis of mandelalide A and its ring-expanded macrolide isomer isomandelalide A has been achieved. Unexpected high levels of cytotoxicity were observed with the ring-expanded isomandelalide A with a rank order of potency: mandelalide A > isomandelalide A > mandelalide B. Key aspects of the synthesis include Ag-catalyzed cyclizations (AgCC's) to construct both the THF and THP rings present in the macrocycle, diastereoselective Sharpless dihydroylation of a cis-enyne, and lithium acetylide coupling with a chiral epoxide.

Full text of DOI:10.1021/jacs.5b12318

Communication
pubs.acs.org/JACS
Enantioselective Total Synthesis of Mandelalide A and
Isomandelalide A: Discovery of a Cytotoxic Ring-Expanded Isomer
Nagarathanam Veerasamy,† Ankan Ghosh,† Jinming Li,† Kazuhiro Watanabe,†,‡ Jeffrey D. Serrill,§
Jane E. Ishmael,§ Kerry L. McPhail,§ and Rich G. Carter*,†
§
†Department of Chemistry and Department of Pharmaceutical Sciences, Oregon State University, Corvallis, Oregon 97331, United
States
‡Laboratory of Synthetic and Medicinal Chemistry, Tohoku Pharmaceutical University, Sendai 981-8558, Japan
S
* Supporting Information
intriguing strategic premise was the possibility that the
ABSTRACT: The total synthesis of mandelalide A and its
ring-expanded macrolide isomer isomandelalide A has
been achieved. Unexpected high levels of cytotoxicity were
observed with the ring-expanded isomandelalide A with a
rank order of potency: mandelalide A > isomandelalide A
> mandelalide B. Key aspects of the synthesis include Ag-
catalyzed cyclizations (AgCC’s) to construct both the
THF and THP rings present in the macrocycle,
diastereoselective Sharpless dihydroylation of a cis-enyne,
and lithium acetylide coupling with a chiral epoxide.
macrocyclization event could be conducted with both the C23
and C24 alcohols exposed and governed by strong, preferential
translactonization to the desired C23 macrolactone. Support for
this premise can be seen in the observations that trans-
lactonization was operable with the originally proposed
structure 14 as well as with other macrolides such as
amphidinolides G/H;8 however, the revised structure 2 exists
exclusively as the C23 macrolactone. If this hypothesis proved
problematic, the C24 position could be masked prior to
macrolactonization.
Synthesis of the northern C15−C24 fragment 4 commenced
from the alkyne 96b,d and iodide 109 (Scheme 2). These
building blocks were readily available from (S)-malic acid (8).
Sonogashira coupling provided the cis-enyne in excellent yield
using diisopropylamine as the solvent. Sharpless asymmetric
dihydroxylation of cis-alkenes10 has seen limited use in total
synthesis11 despite the widespread utility and prevalence of
anti-diols. We were pleased to observe that Sharpless
dihydroxylation using commercially available (DHQD)2PHAL
provided the desired diastereomer 12 in good yield and
reasonable diastereoselectivity (76%, 3.5:1 dr). Use of the
pseudo-enantiomeric ligand (DHQ)2PHAL gave the opposite
diastereomer in similar selectivity (1:3.6 dr). Interestingly, the
Sharpless ligand specifically designed for cis-alkenes (DHQD-
IND)10 proved less effective (2.2:1 dr). Bisesterification using
benzoyl chloride followed by desilylation generated the AgCC
precursor. While our laboratory has explored the utility of diols
as viable nucleophiles in the AgCC process,6b,d the alternate
combination of using dibenzoates (e.g., 6) has not been
reported. We were pleased to discover that this substrate
rapidly and smoothly undergoes cyclization to provide the
unstable enol benzoate 13. Subsequent in situ treatment with
MeLi·LiBr revealed ketone 14 which also proved unstable until
protection as its silyl ether 15 in good overall yield with
excellent stereoselectivity (68%, > 10:1 dr). Ketone 15 was next
olefinated with the Petasis reagent followed by diastereose-
lective reduction using Rh/Al2O3 to provide the 18S stereo-
chemistry. Finally, removal of the Piv ester followed by
conversion to Wittig salt 4 proceeded smoothly.
n 2012, McPhail et al. reported the isolation of a series of
novel macrolides in small quantities from a new species of
I
Lissoclinum ascidian found in Algoa Bay, South Africa.1 These
compounds were named mandelalides A−D, and their
structures were proposed by extensive spectroscopic analysis.
Given the small quantities of material obtained from isolation,
only preliminary evaluation of their biological properties could
be performed. Mandelalides A and B showed low nanomolar
cytotoxicity to lung (NCI-H460) and neuroblastoma (Neuro-
2A) cancer cell lines. Not surprisingly, these structurally
complex, cytotoxic compounds have attracted considerable
synthetic attention from multiple laboratories. In 2014,
Willwacher and Furstner reported the total synthesis of the
̈
proposed structure 1, which did not match the spectroscopic
data of the natural product.2 Shortly thereafter, Ghosh et al.
reported the synthesis of the agylcone of the proposed structure
of mandelalide A.3 Ye et al. synthesized a series of structural
variants of mandelalide A, which allowed them to unequivocally
establish the absolute structure of mandelalide A as compound
2.4,5 Given the scarcity of 2 and our laboratory’s long-standing
interest in the synthesis of cytotoxic macrolides,6 we sought to
develop a robust synthesis of compound 2 to enable additional
biological evaluation. Herein, we report the enantioselective
total synthesis of mandelalide A (2).
As shown in Scheme 1, our synthesis strategy builds on the
observation that both the THF and the THP rings present in
macrolide 2 could be accessed by a silver-catalyzed cylization
(AgCC) from the requisite propargylic benzoates. Our
laboratory had previously demonstrated the utility of this
general concept in our synthesis of amphidinolides C and F;6b,d
however, we are unaware of any examples for accessing six-
membered oxygenated heterocycles using AgCC.7 Another
Received: November 24, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b12318
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Scheme 1. Retrosynthetic Analysis of Mandelalide A (2)
Scheme 2. Synthesis of Northern Fragment 4
Scheme 3. Synthesis of Southern Fragment 5
that the transformation proceeded smoothly to yield 22 in a 6:1
dr at C5 based on crude 1H NMR. This level of
diasteroselectivity may be explained by the modest level of
enantiopurity (7:1 er) in the alkyne precursor 21. The crude
benzoate was immediately submitted to hydrolysis conditions
(NaOMe, MeOH, rt) to generate the desired C7 ketone 23 in
good yield with no loss in stereochemical integrity. It is
important to note that alternate protocols to convert the enol
benzoate 22 into the ketone 23 (including MeLi·LiBr)6b,d led
to inferior yields and/or reduced dr. Removal of the C12 silyl
ether followed by oxidation and Wittig olefination generated
the ester 24. Borohydride reduction followed by Schmidt
glycosidation5 produced the target glycoside 26 in good yield
and reasonable dr at C1′. Finally, conversion of the C14 ester
into its corresponding aldehyde and cross metathesis13
generated the southern fragment 5.
With the northern phosphonium salt in hand, we shifted our
focus to the southern C1−C14 subunit (Scheme 3). Starting
from the readily available epoxide 20 and alkyne 21 (each
accessible in 5 steps from commercially available sources), BF3·
Et2O-mediated coupling under optimized conditions provided
the homopropargylic alcohol 7. While examples exist of alkyne/
epoxide couplings in the presence of an ester,12 propargylic
benzoates have not been used before in the process. With the
cyclization precursor 7 in hand, we explored the key AgCC to
access the pyran 22. While we are unaware of prior examples of
related silver-catalyzed cyclizations for extending the AgCC
process to larger ring systems,7 we were gratified to discover
Our first generation approach to coupling and macro-
cyclization is shown in Scheme 4. Treatment of phosphonium
salt 4 with NaHMDS followed by addition of the aldehyde 5
B
DOI: 10.1021/jacs.5b12318
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Scheme 4. Total Synthesis of Isomandelalide A (31)
Scheme 5. Total Synthesis of Mandelalide A (2)
Table 1. Cytotoxic Effects of Compounds 2, 31, and 33
Against Human Cancer Cell Lines
HeLa cervix
H460 lung
compd. no. EC50 (nM) 95% C.I. (nM) EC50 (nM) 95% C.I. (nM)
2
10.5
17.1
35.5
8.6−12.2
13.1−21.7
16.1−52.4
13.2
30.3
35.5
8.6−18.0
15.0−42.5
26.3−61.2
cleanly provided the C12−C15 E,Z-diene. Next, selective
deprotection of the C23−C24 acetonide in the presence of the
C7 glycoside and the acid-sensitive E,Z-diene motif was
accomplished using TFA to generate the C23,24 diol in excellent
yield. Yamaguchi conditions smoothly induced macrolactoniza-
tion; however, the ring-expanded, 25-membered isomer 29 was
the primary isolable product from this reaction. Removal of the
three silyl moieties in 29 using TAS-F provided the previously
unknown, 25-membered macrolide called isomandelalide A
(31). A second product 30 was also observed in the
macrolactonization. Subsequent TAS-F treatment of the
crude mixture from the macrolactonization revealed the desired
mandelalide A (2); however, only minor amounts of 2 could be
accessed through this route. The preference for 29 over 30 is
surprising based on the observation by Ye et al. that 2 appeared
to prefer the smaller, 24-membered macrolide.4,14
In order to selectively access the desired natural product, the
C24 position was masked as its TBS ether on diol 27 (Scheme
5). Subsequent conversion of the methyl ester into the acid 32
was accomplished by a redox process. Attempted direct
saponification of the C1 methyl ester on the C24 silylated
product led to extensive silyl migration. We were pleased to see
that Yamaguchi macrolactonization of the seco acid 32 proved
effective on the C24 silyl ether series to give the known Ye
intermediate.4 Finally, global desilylation generated the natural
product, the data for which matched with the literature values
(1H, 13C, [α]D).1 A modest amount of a second macrolide was
observed under the cyclization process, which we hypothesize
to be the C21 macrolactone (likely due to C21/C23 silyl
migration); however, this minor product decomposed under
the TAS-F conditions.
31
33
bioassay-guided drug discovery screen,1 synthetic 2 showed
biological activity that was entirely consistent with the
cytotoxicity observed in earlier testing of the natural product.
Synthetic 2 was approximately 3-fold more potent than 33
against human HeLa cervical and H460 lung cancer cells.
Although slightly less potent than 2, isomandelalide A (31) was
also a fully efficacious cytotoxin against both human cancer cell
types with EC50 values of 17.1 and 30.3 nM for HeLa and NCI-
H460 cells, respectively.
In conclusion, efficient total syntheses of mandelalide A (2)
and its ring-expanded isomer isomandelalide A (31) have been
completed. The unexpected potency of isomandelalide A raises
intriguing questions about the biology of these natural
products, given that variation in ring size of a macrocycle
often leads to conformational changes that can have a notable
impact on biological activity (e.g., amphidinolide G/H15 and
isoapoptolidin/apoptolidin16). The disclosed synthetic cam-
paign demonstrated the utility of the AgCC for the
construction of oxygenated heterocycles, including the first
reported example for accessing a pyran ring system. The E,Z-
diene motif was incorporated smoothly through a Wittig
reaction. The macrocyclization preferences for the C23,24 diol
were explored. Further application of this methodology to
access other members of the mandelalide family will be
disclosed in due course.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.5b12318.
Isomandelalide A (31) showed potent cytotoxicity in
comparative testing with synthetic mandelalide A (2) and
relative to the natural mandelalide B (33) (Table 1). Using
assay conditions comparable to those used in the original
■
S
C
DOI: 10.1021/jacs.5b12318
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
(14) Blakemore, P. R.; Browder, C. C.; Hong, J.; Lincoln, C. M.;
Nagorney, P. A.; Robarge, L. A.; Wardrop, D. J.; White, J. D. J. Org.
Chem. 2005, 70, 5449−5460.
(15) Kobayashi, J.; Shigemori, H.; Ishibashi, M.; Yamasu, T.; Hirota,
H.; Sasaki, T. J. Org. Chem. 1991, 56, 5221−5224.
(16) Wender, P. A.; Gulledge, A. V.; Jankowski, O. D.; Seto, H. Org.
Lett. 2002, 4, 3819−3822.
Complete experimental details and biological protocols
(PDF)
1H and 13C NMR spectra (PDF)
AUTHOR INFORMATION
Corresponding Author
*rich.carter@oregonstate.edu
■
Funding
No competing financial interests have been declared.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
Financial support provided by Oregon State University.
Tohuku Pharmaceutical University is acknowledged for their
■
́
fellowship for K.W. Prof. Claudia Maier and Jeff Morre (OSU)
are acknowledged for mass spectra. Dr. Mohamed F. El-Mansy
(OSU) is acknowledged for his assistance with organizing the
spectral data. Finally, the authors are grateful to Prof. Paul Ha-
Yeon Cheong (OSU) and Dr. Roger Hanselmann (Akebia
Therapeutics) for their helpful discussions.
REFERENCES
■
(1) (a) Sikorska, J.; Hau, A. M.; Anklin, C.; Parker-Nance, S.; Davies-
Coleman, M. T.; Ishmael, J. E.; McPhail, K. L. J. Org. Chem. 2012, 77,
6066−6075. (b) Sikorska, J.; Hau, A. M.; Anklin, C.; Parker-Nance, S.;
Davies-Coleman, M. T.; Ishmael, J. E.; McPhail, K. L. J. Org. Chem.
2013, 78, 2812. (c) Davies-Coleman, M. T.; Veale, C. G. L. Mar. Drugs
2015, 13, 6366−6383.
(2) (a) Willwacher, J.; Furstner, A. Angew. Chem., Int. Ed. 2014, 53,
̈
4217−4221. See also: (b) Furstner, A. Angew. Chem., Int. Ed. 2014, 53,
̈
8587−8598.
(3) Reddy, K. M.; Yamini, V.; Singarapu, K. K.; Ghosh, S. Org. Lett.
2014, 16, 2658−2660.
(4) Lei, H.; Yan, J.; Yu, J.; Liu, Y.; Wang, Z.; Xu, Z.; Ye, T. Angew.
Chem., Int. Ed. 2014, 53, 6533−6537.
(5) (a) Wilwacher, J.; Heggen, D.-I. B.; Wirtz, C.; Thiel, W.; Furtsner,
̈
A. Chem. - Eur. J. 2015, 21, 10416−10430. (c) Brutsch, T. M.; Bucher,
̈
P.; Altmann, K.-H. Eur. J. Chem. DOI: 10.1002/chem.201504918.
(6) (a) Lu, L.; Zhang, W.; Carter, R. G. J. Am. Chem. Soc. 2008, 130,
7253−7255. (b) Mahapatra, S.; Carter, R. G. Angew. Chem., Int. Ed.
2012, 51, 7948−7951. (c) Lu, L.; Zhang, W.; Nam, S.; Horne, D. A.;
Jove, R.; Carter, R. G. J. Org. Chem. 2013, 78, 2213−2247.
(d) Mahapatra, S.; Carter, R. G. J. Am. Chem. Soc. 2013, 135,
10792−10803.
(7) (a) Gockel, B.; Krause, N. Org. Lett. 2006, 8, 4485−4488.
(b) Wang, Y.-M.; Kuzniewski, C. N.; Rauniyar, V.; Hoong, C.; Toste,
F. D. J. Am. Chem. Soc. 2011, 133, 12972−12975. (c) Kotikalapudi, R.;
Kumara Swamy, K. C. Tetrahedron 2013, 69, 8002−8012.
(8) Kobayashi, J.; Shimbo, K.; Sato, M.; Tsuda, M. J. Org. Chem.
2002, 67, 6585−6592.
(9) Taguchi, T.; Kitagawa, O.; Morikawa, T.; Nishiwaki, T.; Uehara,
H.; Endo, H.; Kobayashi, Y. Tetrahedron Lett. 1986, 27, 6103−6106.
(10) Wang, L.; Sharpless, K. B. J. Am. Chem. Soc. 1992, 114, 7568−
7570.
(11) (a) Dong, S.; Paquette, L. A. J. Org. Chem. 2005, 70, 1580−
1596. (b) Halim, R.; Brimble, M. A.; Merten, J. Org. Lett. 2005, 7,
2659−2662. (c) Hicks, J. D.; Flamme, E. M.; Roush, W. R. Org. Lett.
2005, 7, 5509−5512. (d) Ermolenko, L.; Sasaki, N. A. J. Org. Chem.
2006, 71, 693−703. (e) Yuen, T.-Y.; Brimble, M. A. Org. Lett. 2012,
14, 5154−5157.
(12) (a) Ramana, C. V.; Raghupathi, N.; Gujar, M. K.; Chorghade,
M. S. Tetrahedron Lett. 2005, 23, 4073−4075. (b) Tinsley, J. M.;
Roush, W. R. J. Am. Chem. Soc. 2005, 127, 10818−10819.
(13) Carlson, E. C.; Rathbone, L. K.; Yang, H.; Collett, N. D.; Carter,
R. G. J. Org. Chem. 2008, 73, 5155−5158.
D
DOI: 10.1021/jacs.5b12318
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX