Second-generation synthesis of azadirachtin: A concise preparation of the propargylic mesylate fragment

Article abstract of DOI:10.1002/anie.200805395

A second bite of the apple: A new and highly efficient synthesis of the propargylic mesylate fragment of azadirachtin has been accomplished (see scheme; Bn=benzyl, Ms=methanesulfonyl, PMB=paramethoxybenzyl, TBDPS=tert- butyldiphenylsilyl). An enantioselec

Full text of DOI:10.1002/anie.200805395

Angewandte
Chemie
DOI: 10.1002/anie.200805395
Natural Product Synthesis
Second-Generation Synthesis of Azadirachtin: A Concise Preparation
of the Propargylic Mesylate Fragment**
Alistair Boyer, Gemma E. Veitch, Edith Beckmann, and Steven V. Ley*
Azadirachtin (1) is a complex natural
product which was first isolated from the
Indian neem tree in 1968 (Figure 1).[1] It is
a potent insect antifeedant and growth
inhibitor;[2] and this, combined with its
challenging architecture, has generated
great interest in its chemical synthesis.[3–5]
Recently,[4] we reported the first syn-
thesis of azadirachtin by using an approach
which evolved over several iterations
(Scheme 1).[7] The most significant consid-
eration in our approach was the formation
ꢀ
of the C8 C14 bond: it was demonstrated
that the O-alkylation of decalin ketone 3
with a propargylic mesylate 2 and subse-
Scheme 1. Strategy for the synthesis of azadirachtin (1).[3,4,7] Bn=benzyl, Ms=methane-
sulfonyl, PMB=para-methoxybenzyl, TBS=tert-butyldimethylsilyl, TES=trimethylsilyl.
turn could arise through an oxidative sequence from the enol
ether 8. The stereogenic center at C17 could be formed by the
diastereoselective reduction of the dihydropyranone 9, which
should be available from an enantioselective hetero Diels–
Alder reaction.[9]
Firstly, we investigated the hetero Diels–Alder reaction,
for which the aldehyde component 12 was synthesized from
readily available 2-propyn-1,4-diol (Scheme 3).[10] However,
Figure 1. Structure of azadirachtin (1).[6]
quent Claisen rearrangement can form this hindered central
linkage. The resulting allene then served as the terminus for a
C17-centered radical, thus allowing rapid access to the desired
tricyclic framework (4!5) and ultimately the natural product
1.
there are few examples of alkynal derivatives used as
substrates in the hetero Diels–Alder reaction,[11] and further-
more the aldehyde 12 was highly unstable. As a result, many
catalysts used in hetero Diels–Alder reactions were unsuit-
able for this transformation and gave either low yield or poor
Although the original route provided sufficient quantities
of the propargylic mesylate 2 to complete the synthesis, the
route was unnecessarily long. Herein, we describe a concise
second-generation approach by building on the knowledge
gained during earlier studies.
It was envisaged that the tetrahydrofuran ring present in 2
could be formed by the selective ozonolysis and methanolysis
of a hemiacetal 6 (Scheme 2).[8] The requisite allyl group
could be introduced by the allylation of a ketone 7, which in
[*] Dr. A. Boyer, Dr. G. E. Veitch, Dr. E. Beckmann, Prof. Dr. S. V. Ley
Department of Chemistry
University of Cambridge
Lensfield Road, Cambridge, CB2 1EW (UK)
Fax: (+44)1223 336442
E-mail: svl1000@cam.ac.uk
[**] We gratefully acknowledge the EPSRC (A.B. and G.E.V.) and the
Alexander von Humboldt Foundation (E.B.) for funding.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200805395.
Scheme 2. Second-generation synthesis plan. PG=protecting group,
TBDPS=tert-butyldiphenylsilyl.
Angew. Chem. Int. Ed. 2009, 48, 1317 –1320
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1317

Communications
Scheme 4. Undesired stereochemical outcome in the allylation of the
a-keto lactone 16: a) OsO4, N-methylmorpholine N-oxide, tBuOH/H2O
(9:1), RT, 98%; b) TEMPO, nBu4NCl, KBr, NaOCl, CH2Cl2/water/brine
(2:1:1), 08C, 71%; c) Dess–Martin periodinane, CH2Cl2, RT; then
toluene, concentrate in vacuo; d) (allyl)nBu3Sn, ZnCl2, CH2Cl2, RT, 54%
(over 2 steps). TEMPO=2,2,6,6-tetramethyl-piperidin-1-oxyl.
Scheme 3. Preparation of enol ether 14. a) 10 (2.5 mol%), CH2Cl2,
ꢀ608C; then TFA, RT, 90% ee, 77% (over 2 steps); b) ZnCl2, CH2Cl2,
08C to RT, 86% (over 2 steps); c) NaBH4, CeCl3·7H2O, MeOH, EtOH,
ꢀ115 to ꢀ908C, 83%; d) Ac2O, NEt3, CH2Cl2, RT, 96%; e) PS “Amano”
lipase (Burklodleria cepacia), aqueous pH 7 buffer/acetone (9:1), RT,
43% (45% recovered acetate); f) TBSCl, imidazole, CH2Cl2, RT, 95%.
TMS=trimethylsilyl, TFA=trifluoroacetic acid.
of the enol ether 14 proceeded smoothly, but attempts to
directly oxidize the corresponding diol to the dicarbonyl 16
proved fruitless. Instead, a stepwise sequence was used in
which first, oxidation with TEMPO led to the a-hydroxy
lactone 15. The second oxidation, to generate the dicarbonyl
16, proved much more challenging. Analysis of the reaction
mixture indicated that treatment of the alcohol 15 with Dess–
Martin periodinane[15] generated the desired product 16, but
attempts to isolate it by using the standard work-up protocol
resulted in decomposition. Attempts to effect in situ allyla-
tion of dicarbonyl 16 were similarly unsuccessful. After these
disappointing results, we tried to find a method of isolating
the fragile product 16, but both direct concentration of the
reaction mixture and attempts to neutralize the reaction
mixture resulted in decomposition of the labile product.
Eventually, we were able to access the desired dicarbonyl 16
(as a mixture with the oxidant by-product: benziodoxolone
17) by adding toluene to the crude reaction mixture and
removing the acid by-product as a toluene/acetic acid azeo-
trope.[16] With the dicarbonyl 16 in hand, the allylation was
readily achieved using allyl tri-n-butyl stannane and zinc(II)
chloride.[17] Although the problem of creating and handling
the sensitive dicarbonyl 16 had been solved, the selectivity of
the allylation reaction gave exclusively the undesired C20-
epimer 18! To explain this result, we propose that the
molecule adopts a half-chair conformation in which axial
attack promotes delivery of the allyl group from the same side
as all of the other substituents on the ring.
enantiodiscrimination. After considerable experimentation,
Hashimotoꢀs dirhodium carboximidate catalyst 10[11b]—a
catalyst specifically optimized for the hetero Diels–Alder
reaction of propargylic aldehydes—was able to promote the
desired cycloaddition, and gave 90% ee and 77% yield over
the two steps. Although this approach represents a highly
efficient preparation of the hetero Diels–Alder adduct 9, the
cost of the catalyst and time-consuming nature of its
preparation encouraged us to prepare 9 as a racemate with
the aim of resolving the two enantiomers at a later stage in the
synthesis.
With sufficient quantities of the dihydropyranone 9 now
available, we next examined the reduction of the ketone at
C17. Under Luche conditions[12] at ꢀ788C the desired
syn geometry (13) was formed with an 8:1 preference, but
by simply lowering the temperature to ꢀ1158C this ratio
increased to 11:1. Pleasingly, the undesired epimer could be
readily separated and recycled to give the starting ketone 9 by
oxidation with TPAP.[13] It was at this stage that we opted to
resolve the racemic intermediate and thus the allylic alcohol
rac-13 was converted into the corresponding acetate. We
found that the PS “Amano” lipase worked rapidly and
selectively,[14] and gave the product 13 as a single enantiomer.
It was anticipated that the secondary alcohol 13 could be
readily protected as its PMB ether (see 2), but all attempts to
install the PMB protecting group met with failure. Instead,
protection of 13 as its TBS ether proved straightforward and
gave the enol ether 14.
However, the enol ether 14 offered several alternatives for
the installation of the remaining functionality in the mesylate
2 and we next investigated an acetal unit at C21 (19;
Scheme 5). Treatment of the enol ether 14 with DMDO
rapidly and selectively generated the glycal oxide which, in
methanol, underwent ring opening to provide the methyl
At this stage we faced a number of possibilities with which
to introduce the tetrahydrofuran ring in 2 and we initially
targeted the a-keto lactone 16 (Scheme 4). Dihydroxylation
1318
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 1317 –1320

Angewandte
Chemie
the hemiacetal 23. Then, the lactol was transformed into the
desired methyl acetal 24 (as a mixture of epimers at C23) after
acidic methanolysis. It turned out to be serendipitous that
earlier attempts to introduce a PMB ether group at C17 were
unsuccessful, as its use to protect the hemiacetal unit at C21
was key in the synthesis of the core of mesylate 2.
With the carbon framework of the mesylate 2 in place, all
that remained was manipulation of the protecting groups
(Scheme 7). Treatment of the bis(silyl ether) 24 with one
Scheme 5. Preparation of methyl acetal 20 for the synthesis of 2:
a) DMDO, acetone/CH2Cl2 (1:1), 08C; then MeOH, 82%; b) Dess–
Martin periodinane, CH2Cl2, RT; c) (allyl)MgBr, THF, ꢀ788C, 78%
(over 2 steps); d) NaH, BnBr, DMF, 08C to RT, 62%. DMDO=di-
methyldioxirane, DMF=N,N-dimethylformamide, THF=tetrahydro-
furan.
acetal 19. Standard manipulations then led to the intermedi-
ate 20, which possesses much of the functionality required to
complete the synthesis. Nevertheless, reaction conditions to
selectively cleave the methyl acetal at C21 in 20 could not be
found.[18]
Therefore, we chose an alternative acetal derivative. The
glycal oxide (from 14) was treated with PMB alcohol and
zinc(II) chloride to form the PMB-protected acetal 21
(Scheme 6). The nonpolar nature of the molecule allowed
Scheme 7. Final steps in the second-generation synthesis of propar-
gylic mesylate fragments 2 and 26: a) TBAF, THF, RT, 39% (over 4
steps); b) TBAF, THF, RT, 39% (over 4 steps); c) TBDPSCl, imidazole,
CH2Cl2, RT, 87%; d) NaH, PMBBr, nBu4NI, DMF, 08C to RT, 26%
a-29, 52% b-29; e) TBAF, THF, RT, 96% a-30; 84% b-30; f) Ms2O,
iPr2NEt, CH2Cl2, 08C, 90%. TBAF=tetra-n-butylammonium fluoride.
equivalent of either HF·pyridine or TBAF resulted in highly
selective formation of the propargyl alcohol 25. The corre-
sponding propargylic mesylate 26 represents a potential
intermediate for a second-generation synthesis of azadirach-
tin; however, the fragment 2 that was used previously in the
synthesis of azadirachtin contained a PMB ether at C17.
Consequently, the bis(silyl ether) 24 was treated with an
excess of TBAF to effect complete desilylation. Reprotection
of the less hindered primary alcohol 27 allowed access to the
secondary alcohol 28 which could then be transformed into
the corresponding PMB ether 29 using PMB bromide and
sodium hydride in the presence of nBu4NI. At this stage the
C23 epimers 29 were separated by column chromatography
and, as both represent precursors to the natural product,[4a]
TBAF was used to effect removal of the TBDPS protecting
group to give the corresponding propargylic alcohols a-30/b-
30, which were identical in all respects to those prepared
previously[7] (see the Supporting Information).
In summary, by generating and using the a-keto lactone 16
we have discovered a highly selective method of introducing
substitution in a d-lactone and have devised a convenient
work-up procedure for the isolation of sensitive Dess–Martin
oxidation products. After extensive experimentation, it was
found that the PMB-substituted acetal 22 could be elaborated
to complete a second-generation synthesis of propargylic
mesylate 2. Notably, all of the stereogenic centers in the
Scheme 6. Completion of the core for propargylic mesylate 2:
a) DMDO, acetone/CH2Cl2 (1:1), 08C; then PMBOH, ZnCl2, CH2Cl2,
08C to RT, 86%; b) Dess–Martin periodinane, CH2Cl2, RT;
c) (allyl)MgBr, THF, ꢀ1108C; d) NaH, BnBr, DMF, 08C to RT, 52%
(over 3 steps); e) O3, CH2Cl2, ꢀ788C; then PS–PPh3, RT; f) CH2Cl2
(90 ppm water)/TFA (9:1), RT; g) Amberlyst A-15, molecular sieves
(3 ꢀ), MeOH/MeCN (1:10), RT, 1:1 to 1:5 (a/b) mixture of C23-
epimers. PS=polystyrene supported.
the excess PMB alcohol to be removed simply by aqueous
extraction. Oxidation with the Dess–Martin reagent[15] pro-
ceeded smoothly to access the ketone at C20. Then, addition
of allyl Grignard reagent at ꢀ1108C generated the desired
configuration at C20 with excellent selectivity, and the
resulting tertiary alcohol was protected as its benzyl ether
22—as required for the synthesis of 2. When we attempted to
prepare the hemiacetal (see 6; Scheme 2), extensive decom-
position of the substrate 22 was observed upon treatment with
DDQ. The use of trifluoroacetic acid solved this problem[19]
and promoted the desired cleavage of the acetal unit to give
the d-lactol. However, we observed an improvement in the
overall yield if the sequence of events was altered. Ozonolysis
of 22 smoothly generated the aldehyde, which was treated
with TFA to effect cleavage of the PMB acetal and resulted in
Angew. Chem. Int. Ed. 2009, 48, 1317 –1320
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1319

Communications
2005, 44, 3447; g) H. Watanabe, N. Mori, D. Itoh, T. Kitahara, K.
Mori, Angew. Chem. 2007, 119, 1534; Angew. Chem. Int. Ed.
2007, 46, 1512.
molecule were selectively installed using the single stereo-
center at C15, which was set by a highly effective enantiose-
lective hetero Diels–Alder reaction. Overall, this route
represents a highly efficient preparation of the azadirachtin
coupling partner 2, and proceeded in only 17 steps (compared
to 26 steps for the original route). Studies are currently
underway towards a second-generation synthesis of the
decalin portion 3 of this ever-fascinating molecule azadir-
achtin (1).
[6] a) C. J. Turner, M. S. Tempesta, R. B. Taylor, M. G. Zagorski,
J. S. Termini, D. R. Schroeder, K. Nakanishi, Tetrahedron 1987,
43, 2789; b) J. N. Bilton, H. B. Broughton, P. S. Jones, S. V. Ley,
Z. Lidert, E. D. Morgan, H. S. Rzepa, R. N. Sheppard, A. M. Z.
Slawin, D. J. Williams, Tetrahedron 1987, 43, 2805; c) W. Kraus,
M. Bokel, A. Bruhn, R. Cramer, I. Klaiber, A. Klenk, G. Nagl,
H. Pꢁhnl, H. Sadlo, B. Vogler, Tetrahedron 1987, 43, 2817.
[7] For full details of the synthesis of azadirachtin, see: S. V. Ley,
et al., Chem. Eur. J. 2008, 14, 10683.
Received: November 4, 2008
[8] For strategies used to construct the furanacetal portion of
azadirachtin, see: a) S. V. Ley, D. Santafianos, W. M. Blaney,
M. S. J. Simmonds, Tetrahedron Lett. 1987, 28, 221; b) D.
Pflieger, B. Muckensturm, P. C. Robert, M. T. Simonis, J. C.
Kienlen, Tetrahedron Lett. 1987, 28, 1519; c) Y. Nishikimi, T.
Iimori, M. Sodeoka, M. Shibasaki, J. Org. Chem. 1989, 54, 3354;
d) J. C. Anderson, S. V. Ley, Tetrahedron Lett. 1990, 31, 431;
e) K. J. Henry, B. Fraser-Reid, J. Org. Chem. 1994, 59, 5128; f) H.
Watanabe, T. Watanabe, K. Mori, Tetrahedron 1996, 52, 13939;
g) J. Ishihara, T. Fukuzaki, A. Murai, Tetrahedron Lett. 1999, 40,
1907; h) S. V. Ley, Pure Appl. Chem. 2005, 77, 1115.
Published online: January 12, 2009
Keywords: a-keto lactones · asymmetric catalysis · hetero Diels–
.
Alder reaction · natural products · total synthesis
[1] J. H. Butterworth, E. D. Morgan, Chem. Commun. 1968, 23.
[2] For reviews, see: a) A. J. Mordue, E. D. Morgan, A. J. Nisbet in
Comprehensive Molecular Insect Science: Control, Vol. 6 (Eds.:
L. I. Gilbert, K. Iatrou, S. S. Gill), Elsevier, Oxford, 2004;
b) A. J. Mordue, M. S. J. Simmonds, S. V. Ley, W. M. Blaney, W.
Mordue, M. Nasiruddin, A. J. Nisbet, Pestic. Sci. 1998, 54, 277;
c) A. J. Mordue, A. Blackwell, J. Insect Physiol. 1993, 39, 903;
d) H. Schmutterer, The Neem Tree, Wiley-VCH, Weinheim,
1995.
[9] S. Danishefsky, Acc. Chem. Res. 1981, 14, 400.
[10] C. F. Morrison, D. J. Burnell, Tetrahedron Lett. 2001, 42, 7367.
[11] We only found five publications describing the hetero Diels–
Alder reaction of propargylic aldehydes: a) S. Kii, T. Hashimoto,
K. Maruoka, Synlett 2002, 931; b) M. Anada, T. Washio, N.
Shimada, S. Kitagaki, M. Nakajima, M. Shiro, S. Hashimoto,
Angew. Chem. 2004, 116, 2719; Angew. Chem. Int. Ed. 2004, 43,
2665; c) A. K. Unni, N. Takenaka, H. Yamamoto, V. H. Rawal, J.
Am. Chem. Soc. 2005, 127, 1336; d) T. Washio, H. Nambu, M.
Anada, S. Hashimoto, Tetrahedron: Asymmetry 2007, 18, 2606;
e) T. Washio, R. Yamaguchi, T. Abe, H. Nambu, M. Anada, S.
Hashimoto, Tetrahedron 2007, 63, 12037.
[3] For a review, see: G. E. Veitch, A. Boyer, S. V. Ley, Angew.
Chem. 2008, 120, 9542; Angew. Chem. Int. Ed. 2008, 47, 9402.
[4] a) G. E. Veitch, E. Beckmann, B. J. Burke, A. Boyer, S. L.
Maslen, S. V. Ley, Angew. Chem. 2007, 119, 7773; Angew. Chem.
Int. Ed. 2007, 46, 7629; b) G. E. Veitch, E. Beckmann, B. J.
Burke, A. Boyer, C. Ayats, S. V. Ley, Angew. Chem. 2007, 119,
7777; Angew. Chem. Int. Ed. 2007, 46, 7633; c) G. E. Veitch, A.
Pinto, A. Boyer, E. Beckmann, J. C. Anderson, S. V. Ley, Org.
Lett. 2008, 10, 569.
[5] For advanced synthesis studies, see: a) H. Watanabe, T. Wata-
nabe, K. Mori, T. Kitahara, Tetrahedron Lett. 1997, 38, 4429;
b) T. Fukuzaki, S. Kobayashi, T. Hibi, Y. Ikuma, J. Ishihara, N.
Kanoh, A. Murai, Org. Lett. 2002, 4, 2877; c) J. Ishihara, Y.
Ikuma, S. Hatakeyama, T. Suzuki, A. Murai, Tetrahedron 2003,
59, 10287; d) K. C. Nicolaou, A. J. Roecker, H. Monenschein, P.
Guntupalli, M. Follmann, Angew. Chem. 2003, 115, 3765; Angew.
Chem. Int. Ed. 2003, 42, 3637; e) K. C. Nicolaou, P. K. Sasmal,
A. J. Roecker, X. W. Sun, S. Mandal, A. Converso, Angew.
Chem. 2005, 117, 3509; Angew. Chem. Int. Ed. 2005, 44, 3443;
f) K. C. Nicolaou, P. K. Sasmal, T. V. Koftis, A. Converso, E.
Loizidou, F. Kaiser, A. J. Roecker, C. C. Dellios, X. W. Sun, G.
Petrovic, Angew. Chem. 2005, 117, 3513; Angew. Chem. Int. Ed.
[12] J. L. Luche, J. Am. Chem. Soc. 1978, 100, 2226.
[13] a) W. P. Griffith, S. V. Ley, G. P. Whitcombe, A. D. White, J.
Chem. Soc. Chem. Commun. 1987, 1625; b) S. V. Ley, J. Norman,
W. P. Griffith, S. P. Marsden, Synthesis 1994, 639.
[14] The enzyme was purchased from Amano Enzyme Inc. For a
similar application, see: K. Sugawara, Y. Imanishi, T.
Hashiyama, Tetrahedron: Asymmetry 2000, 11, 4529.
[15] D. B. Dess, J. C. Martin, J. Org. Chem. 1983, 48, 4155.
[16] L. H. Horsley, Azeotropic Data III, American Chemical Society,
Washington, DC, 1973, p. 112.
[17] The allylation reaction tolerated the presence of 17.
[18] Under a variety of reaction conditions, the TBS then the benzyl
ether groups were cleaved while the methyl acetal unit remained
intact.
[19] L. Yan, D. Kahne, Synlett 1995, 523.
1320
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 1317 –1320