Total synthesis and structural elucidation of azaspiracid-1. Synthesis-based analysis of originally proposed structures and indication of their non-identity to the natural product

Article abstract of DOI:10.1021/ja054748z

The key building blocks (6, 7, and 8) for the intended construction of the originally proposed structures of azaspiracid-1, a potent marine-derived neurotoxin, were coupled and the products elaborated to the targeted compounds (1a,b) and their C-20 epimer

Full text of DOI:10.1021/ja054748z

Published on Web 01/28/2006
Total Synthesis and Structural Elucidation of Azaspiracid-1.
Synthesis-Based Analysis of Originally Proposed Structures
and Indication of Their Non-Identity to the Natural Product
K. C. Nicolaou,* David Y.-K. Chen, Yiwei Li, Noriaki Uesaka, Goran Petrovic,
Theocharis V. Koftis, Federico Bernal, Michael O. Frederick, Mugesh Govindasamy,
Taotao Ling, Petri M. Pihko, Wenjun Tang, and Stepan Vyskocil
Contribution from the Department of Chemistry and The Skaggs Institute for Chemical Biology,
The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037,
and Department of Chemistry and Biochemistry, UniVersity of California, San Diego,
9500 Gilman DriVe, La Jolla, California 92093
Received July 15, 2005; E-mail: kcn@scripps.edu
Abstract: The key building blocks (6, 7, and 8) for the intended construction of the originally proposed
structures of azaspiracid-1, a potent marine-derived neurotoxin, were coupled and the products elaborated
to the targeted compounds (1a,b) and their C-20 epimers (2 and 3). The assembly of the three intermediates
was accomplished by a dithiane-based coupling reaction that united the C1-C20 (7) and C21-C27 (8)
fragments, followed by a Stille-type coupling which allowed the incorporation of the C28-C40 fragment (6)
into the growing substrate. Neither of the final products (1a,b) matched the natural substance by TLC or
1H NMR spectroscopic analysis, suggesting one or more errors in the originally proposed structure for this
notorious biotoxin.
Introduction
Results and Discussion
In the preceding paper,1 we described the stereoselective
construction of the three key building blocks 6, 7, and 8 (Figure
3) required for our projected total synthesis of the proposed
structure of azaspiracid-1 (1a, Figure 1).2 These sequences
delivered all three fragments in both enantiomeric forms as
needed to ensure the eventual assignment of the relative and
absolute stereochemistry of the natural product. In this article
we describe the coupling of these key building blocks and the
elaboration of the resulting products to the targeted molecules,
the two diastereomeric azaspiracid-1 structures 1a and 1b
(Figure 1). This accomplishment, however, only proved that
the originally proposed structure (1a or 1b) was in error. This
finding prompted the synthesis of the two C-20-epi diastereo-
mers 2 and 3 (Figure 1), neither of which matched the natural
product, thereby proving their non-identity to the true structure
of azaspiracid-1. This, in turn, left the project open for new
speculations and initiatives for the deconvolution of the still
remaining puzzle of this intriguing natural product. The de-
mystification of the true structure of azaspiracid-1 as that
depicted by 1 (Figure 2) will be described in a subsequent
publication.3
1. Coupling of the C1-C20 and C21-C27 Fragments and
Synthesis of the ABCDE Domain. Of the two available choices
to assemble the C1-C20 (7), C21-C27 (8), and C28-C40 (6)
fragments into the targeted azaspiracid-1 architecture, we opted
for that involving initial union of 7 and 8 through a dithiane
coupling,4 followed by subsequent incorporation of 6 into the
growing molecule (i.e., 5) through a Stille reaction,5 as outlined
retrosynthetically in Figure 3. As a prelude to the dithiane-based
coupling of 7 and 8, we carried out a preliminary investigation
involving the simpler and, therefore, more plentiful partners
dithiane 9 and carbonyl compounds 10, 11, 11a, and 11b (Table
1) in order to develop the necessary technology for what was
expected to be a challenging task. Table 1 summarizes the results
of this study. Thus, employing at first the aldehyde substrate
corresponding to structure 10 and t-BuLi or n-BuLi as the base
to form the lithio derivative of dithiane 9 in various solvents
and different temperatures led to no product containing both
fragments. Monitoring the formation of the expected lithiated
compound from 9 by D2O quenching/1H NMR spectroscopic
analysis led to the conclusion that the desired intermediate,
although initially formed, was incapable of being properly
trapped with the aldehyde partner (10, entries 1-6, Table 1),
being too short-lived for a useful reaction.
(1) Nicolaou, K. C.; Pihko, P. M.; Bernal, F.; Frederick, M. O.; Qian, W.;
Uesaka, N.; Diedrichs, N.; Hinrichs, J.; Koftis, T. V.; Loizidou, E.; Petrovic,
G.; Rodriquez, M.; Sarlah, D.; Zou, N. J. Am. Chem. Soc. 2006, 128, 2244-
2257.
(2) Satake, M.; Ofuji, K.; Naoki, H.; James, K. J.; Fruey, A.; McMahon, T.;
Silke, J.; Yasumoto, T. J. Am. Chem. Soc. 1998, 120, 9967.
(3) Nicolaou, K. C.; Koftis, T. V.; Vyskocil, S.; Petrovic, G.; Tang, W.;
Frederick, M. O.; Chen, D. Y.-K.; Li, Y.; Ling, T.; Yamada, Y. M. A. J.
Am. Chem. Soc. 2006, 128, in press.
Faced with this predicament, we then employed the n-BuLi-
n-Bu2Mg reagent, which is known to provide a longer-lived
(4) Corey, E. J.; Seebach, D. Angew. Chem., Int. Ed. Engl. 1965, 4, 1075.
(5) (a) Stille, J. K. Angew. Chem., Int. Ed. Engl. 1986, 25, 508. (b) Del Valle,
L.; Stille, J. K.; Hegedus, L. S. J. Org. Chem. 1990, 55, 3019.
9
2258
J. AM. CHEM. SOC. 2006, 128, 2258-2267
10.1021/ja054748z CCC: $33.50 © 2006 American Chemical Society

Synthesis and Structure of Azaspiracid-1
A R T I C L E S
Figure 1. Two of the originally proposed structures of azaspiracid-1, 1a and 1b, and synthesized C-20 epimers 2 and 3.
NaClO2-based oxidation, 82% yield; and (iv) DCC coupling
with pentafluorophenol, 85% yield], as shown in Scheme 1.
With the coupling reaction of fragments 9 and 11 at a
satisfactory level of efficiency, we proceeded to explore the
sequence to the complete model ABCDE ring framework as
shown in Scheme 1. For this investigation, we utilized the
C5-C27 ketone 16 obtained by coupling of the ABCD segment
11 with dithiane 8 according to the n-BuLi-n-Bu2Mg-based
procedure (63% yield). From the many reducing reagents tested,
only DIBAL-H performed well in this reaction, yielding, at -90
°C in CH2Cl2, the desired and anticipated alcohol as a single
stereoisomer in a reaction that also cleaved the pivaloate ester,
leading to compound 17 (55% yield). The lost ester group was
easily re-introduced (PivCl, py, 75% yield) to furnish the next
intermediate, 18, from which the dithiane moiety was re-
moved by the action of PhI(OCOCF3)2 (78% yield).11 The so
generated keto silyl bis-ether (19) was treated with TBAF,
which allowed the liberated trihydroxy ketone to spontan-
eously fold on itself, affording the hydroxy hemiketal 20 in
85% yield. For the purposes of assigning the stereochemistry
of the reduction, this 1,2-diol system was treated with triphos-
gene and pyridine, giving the cyclic carbonate 21, whose
structural rigidity allowed its 1H NMR spectra and NOE studies
to be particularly revealing.12 Indeed, the observed NOEs (see
Scheme 1) confirmed the shown (and desired) stereochemistry
for the C-20 hydroxyl group which was generated a few steps
back by the DIBAL-H reduction of the carbonyl group at that
position.
Figure 2. Revised structure of azaspiracid-1 (1).
organometallic species from dithianes,6 expecting a positive
outcome in the coupling reaction. Indeed, an encouraging
improvement was observed, with yields up to 42% under certain
conditions (entry 7, Table 1), but the coupling product consisted
of a random (ca. 1:1) mixture of the two diastereomeric alcohols.
Having exhausted all hope for further improvement with
aldehyde 10 as the electrophile partner in this reaction, we
adopted a series of alternative electrophilic equivalents, most
notably activated esters. Such partners, we reasoned, could not
only offer advantages such as a more reliable coupling substrate
in this carbon-carbon bond-forming process but also, most
importantly, provide us with an opportunity to control the C-20
stereochemistry through the subsequent reduction of the ex-
pected ketone (in contrast to the aldehyde coupling that
proceeded in a nonstereoselective manner, as mentioned above).
Among the activated esters utilized (e.g., 2-pyridinethiol 11a,7
benzotriazole 11b,8 and pentafluorophenol 11,9 entries 8-10,
Table 1), the pentafluorophenol ester proved the best, leading
to a 63% yield of the desired ketone (entry 10, Table 1). The
required pentafluoro ester 11 was prepared by a four-step
sequence from TBDPS derivative 13 [(i) TBAF-induced desi-
lylation, 88% yield; (ii) Swern oxidation,10 94% yield; (iii)
2. Synthesis of Two of the Diastereomers of the Originally
Proposed Structures of Azaspiracid-1. With the development
of the synthetic technology for the establishment of the ABCDE
domain of azaspiracid-1, the road was now open for advance-
ment of the C1-C20 fragment 221 to the C1-C27 allylic acetate
29 (Scheme 2), as needed for the planned Stille reaction with
stannane 6. To this end, the lithium anion derived from 8 and
n-BuLi-n-Bu2Mg in THF at 25 °C as described above was
(6) Ide, M.; Nakata, M. Bull. Chem. Soc. Jpn. 1999, 72, 2491.
(7) Balasubramanian, T.; Strachan, J.-P.; Boyle, P. D.; Lindsey, J. S. J. Org.
Chem. 2000, 65, 7919.
(8) Katritzky, A. R.; Abdel-Fattah, A. A. A.; Wang, M. J. Org. Chem. 2003,
68, 4932.
(9) Schmidt, U.; Kroner, M.; Griesser, H. Tetrahedron Lett. 1988, 29, 4407.
(10) Omura, K.; Sharma, A. K.; Swern, D. J. Org. Chem. 1976, 41, 957.
(11) Stork, G.; Zhao, K. Tetrahedron Lett. 1989, 30, 287.
(12) Nicolaou, K. C.; Li, Y.; Sugita, K.; Monenschein, H.; Guntupalli, P.;
Mitchell, H. J.; Fylaktakidou, K. C.; Vourloumis, D.; Giannakakou, P.;
O’Brate, A. J. Am. Chem. Soc. 2003, 125, 15443.
9
J. AM. CHEM. SOC. VOL. 128, NO. 7, 2006 2259

A R T I C L E S
Nicolaou et al.
Figure 3. Retrosynthetic analysis of the originally proposed structure of azaspiracid-1 (1a).
Table 1. Optimization of Dithiane Coupling for the Formation of the C20-C21 Bond of Azaspiracid-1
% deuterium
exchangea
entry
(a) reagents and conditions
R
yield (%)b
X
1
2
3
4
5
6
7
8
t-BuLi, HMPA, THF, -78 °C, 1 h
t-BuLi, HMPA, THF, 0 °C, 5 min
t-BuLi, THF, 25 °C, 1 h
t-BuLi, Et2O, 25 °C, 1 h; MgBr2
n-BuLi, NaO-t-Bu, THF, -78 °C, 1 h
n-BuLi, THF, 0 °C, 30 min
H (10)
H (10)
H (10)
H (10)
H (10)
H (10)
H (10)
92
30
77
85
17
50
80
10
0
0
0
H, OH
H, OH
H, OH
H, OH
H, OH
H, OH
H, OH
O
0
0
n-BuLi-n-Bu2Mg, THF, -90 °C, 1 h
n-BuLi-n-Bu2Mg, THF, -90 °C, 1 h
10-42c
0
(11a)
(11b)
9
n-BuLi-n-Bu2Mg, THF, -90 °C, 1 h
n-BuLi-n-Bu2Mg, THF, -90 °C, 1 h
0
O
O
10
63
(11)
a D2O addition after the indicated reaction time. b Addition of 10, 11, 11a, and 11b at -90 °C. c Combined yield of a 1:1 mixture of diastereoisomers.
reacted at -90 °C with pentafluoro ester 7 [prepared from silyl
ether 22 by desilylation with TBAF, Swern oxidation of the
resulting alcohol 23 to the corresponding aldehyde (24), and
further oxidation to the carboxylic acid (25) followed by DCC-
mediated coupling with pentafluorophenol (64% yield over the
four steps)] to afford dithiane ketone 26 in 63% yield, as shown
in Scheme 2. The DIBAL-H reduction of 26 proceeded, as
expected, stereoselectively and in 55% yield, furnishing a single
9
2260 J. AM. CHEM. SOC. VOL. 128, NO. 7, 2006

Synthesis and Structure of Azaspiracid-1
A R T I C L E S
Scheme 1. Model Dithiane Coupling and Structural Confirmation of All Stereocenters in the ABCDE Ring System (C5-C27 Domain)a
a Reagents and conditions: (a) TBAF (1.0 M in THF, 5.0 equiv), THF, 25 °C, 2 h, 88%; (b) (COCl)2 (5.0 equiv), DMSO (11 equiv), CH2Cl2, -78 °C,
30 min; then Et3N (22 equiv), -78 f 0 °C, 94%; (c) NaClO2 (6.0 equiv), NaH2PO4 (6.0 equiv), 2-methyl-2-butene (excess), t-BuOH:H2O (4:1), 25 °C, 1.5
h; (d) pentafluorophenol (1.5 equiv), DCC (2.0 equiv), CH2Cl2, 25 °C, 2.5 h, 78% over two steps; (e) 8 (9.0 equiv), n-BuLi-n-Bu2Mg (1.1 M in hexanes,
6.0 equiv), THF, 0 f 25 °C, 1.5 h; then 35, -90 °C, 15 min, 63%; (f) DIBAL-H (1.0 M in CH2Cl2, 10 equiv), CH2Cl2, -90 °C, 1.5 h, 55%; (g) PivCl (3.0
equiv), py (10 equiv), 0 f 25 °C, 12 h, 75%; (h) PhI(OCOCF3)2 (2.2 equiv), MeCN:pH 7 buffer (4:1), 0 °C, 78%; (i) TBAF (1.0 M in THF, 5.0 equiv),
THF, 25 °C, 16 h, 85%; (j) triphosgene (2.0 equiv), py (15 equiv), CH2Cl2, -78 f 25 °C, 1 h, 54%. Abbreviations: Piv, trimethylacetyl; TBDPS, tert-
butyldiphenylsilyl; TBAF, tetra-n-butylammonium fluoride; DMSO, dimethyl sulfoxide; PFP, pentafluorophenyl; DCC, dicyclohexylcarbodiimide; DIBAL-
H, diisobutylaluminum hydride.
Scheme 2. Synthesis of Stille Coupling Partner 29a
a Reagents and conditions: (a) TBAF (1.0 M in THF, 2.0 equiv), THF, 0 f 25 °C, 3 h, 93%; (b) (COCl)2 (5.0 equiv), DMSO (11 equiv), CH2Cl2, -78
f -60 °C, 2 h; then Et3N (22 equiv), -78 f -30 °C, 1 h, 89%; (c) NaClO2 (4.0 equiv), NaH2PO4 (4.0 equiv), 2-methyl-2-butene (5.0 equiv), t-BuOH:H2O
(5:1), 25 °C, 2 h, 95%; (d) pentafluorophenol (1.2 equiv), DCC (1.5 equiv), CH2Cl2, 25 °C, 2 h, 82%; (e) 8 (9.0 equiv), n-BuLi-n-Bu2Mg (1.1 M in hexanes,
6.0 equiv), THF, 0 f 25 °C, 1.5 h; then 7, -90 °C, 15 min, 63%; (f) DIBAL-H (1.0 M in CH2Cl2, 10 equiv), CH2Cl2, -90 °C, 1.5 h, 55%; (g) TBAF (1.0
M in THF, 5.0 equiv), THF, 25 °C, 16 h, 78%; (h) Ac2O (50 equiv), pyridine:CH2Cl2 (1:1), 0 f 25 °C, 16 h, 84%; (i) PhI(OCOCF3)2 (2.2 equiv), MeCN:pH
7 buffer (4:1), 0 °C, 78%.
9
J. AM. CHEM. SOC. VOL. 128, NO. 7, 2006 2261

A R T I C L E S
Nicolaou et al.
Scheme 3. Stille Coupling of Stannanes 6 and ent-6 with Allylic Acetate 29, and Arrival at Octacyclic Ring Systems 35 and 36a
a Reagents and conditions: (a) Pd2dba3 (0.3 equiv), AsPh3 (0.3 equiv), LiCl (6.0 equiv), i-Pr2NEt (12 equiv); then 6 or ent-6 (0.03 M in THF, syringe
pump addition), NMP, 45 °C, 4 h, 66% for 30, 56% for 31; (b) HF‚py (excess), THF:pyridine (1:1), 0 f 25 °C, 2.5 h; (c) NIS (10 equiv), NaHCO3 (30
equiv), THF, 0 °C, 16 h, 67% for 33, 63% for 34 over two steps; (d) Et3B (1.0 M in hexanes, 3.0 equiv), n-Bu3SnH:toluene (1:2), 0 °C, 5 min, 92% for 35,
94% for 36. Abbreviations: NIS, N-iodosuccinimide; TES, triethylsilyl; Teoc, 2-(trimethylsilyl)ethoxycarbonyl; dba, dibenzylideneacetone; NMP,
N-methylpyrrolidone.
diol with the C-20 hydroxyl group, assumed at this point to be
of the same configuration as that obtained before with the
reduction of 16 (Scheme 1). This assignment was to be
confirmed later on in the synthesis by X-ray crystallographic
analysis of a descendent compound (vide infra). Exposure of
the latter compound (27) to TBAF resulted in the cleavage
of the cyclic silyl bis-ether, generating tetraol 28 (78%
yield), which was selectively converted to triacetate 5 with
Ac2O in the presence of pyridine in 84% yield. Being shielded
more by its surroundings than the other three, the C-20 hy-
droxyl group remained inert under these conditions. The di-
thiane group was then removed from compound 5 by the ac-
acetate 29 with stannane 6 by a Stille coupling reaction. Due
to the complexity and sensitivity of the substrates involved,
success in this process required considerable experimentation.
As shown in Table 2, employment of Pd(PPh3)4 as a catalyst in
conjunction with LiCl13 and i-Pr2NEt in THF at 45 °C led to
no coupling product formation, with only decomposition
observed (entry 1). Switching to Pd2(dba)3 as the catalyst under
the same conditions resulted in coupling (60% combined yield);
however, the product was found to be a mixture of the desired
compound (30) and its C-25 acetoxy epimer (C-25-epi-30), with
30 predominating in ca. 2:1 ratio. Faced with this epimerization
problem, which we attributed to the high reactivity of the
catalyst, we then proceeded to tame its behavior by employing
AsPh3 as an additional ligand in the reaction mixture. It was
reasoned that the bulkiness of this ligand combined with its
11
tion of PhI(OCOCF3)2 to afford, in 78% yield, triacetoxy
ketone 29, representing the required C1-C27 Stille coupling
partner.
The next task toward the total synthesis of the originally
proposed structure of azaspiracid-1 called for the union of allylic
(13) Farina, V.; Krishnamurthy, V.; Scott, W. J. Org. React. 1997, 50, 54.
9
2262 J. AM. CHEM. SOC. VOL. 128, NO. 7, 2006

Synthesis and Structure of Azaspiracid-1
A R T I C L E S
Table 2. Optimization of the Stille Coupling of 29 with 6
entry
reagents and conditions (a)a
temp (
°
C)
time (h)
yield (%) 30:C25-epi-30
1
2
3
4
5
6
10 mol % Pd(PPh3)4, LiCl, i-Pr2NEt
10 mol % Pd(dba)3, LiCl, i-Pr2NEt
10 mol % Pd2(dba)3, 80 mol % AsPh3, LiCl, i-Pr2NEt
10 mol % Pd2(dba)3, 40 mol % AsPh3, LiCl, i-Pr2NEt
10 mol % Pd2(dba)3, 20 mol % AsPh3, LiCl, i-Pr2NEt
10 mol % Pd2(dba)3, 10 mol % AsPh3, LiCl, i-Pr2NEt
45
45
45
45
45
45
15
12
16
8
5
4
0:0
40:20
10:0
50:0
60:0
66:0
a Compound 29 was dissolved in NMP and LiCl and i-Pr2NEt were added. To this mixture was added palladium catalyst and in entries 2, 3, 4, and 5,
ligand additive AsPh3. Compound 6 was added as a solution in THF at the indicated temperature and over the time period listed.
furnished, in 67% overall yield over the two steps, iodoether
33 as a single diastereomer. Iodide 33 pleasantly surprised us
with its nice crystals suitable for X-ray crystallographic analysis,
which confirmed its structure (and those of its predecessor
compounds) as shown (see ORTEP drawing, Figure 4). It should
be noted that the Hg(OAc)2 technology16 that was successfully
employed to forge this key ring (G) in a model system1 led to
decomposition when applied to substrate 4.
Having accomplished its mission of forging ring G and
proving beyond doubt the configuration of the growing mol-
ecule, the iodide residue was then ejected reductively from
intermediate 33 by exposure to excess n-Bu3SnH and Et3B17 as
a radical initiator in toluene at 0 °C, to afford the octacyclic
system 35 in 92% yield. In an analogous manner and in similar
yields, the diastereomeric compound 36 was synthesized from
Figure 4. ORTEP drawing of iodide 33.
intermediate 29 and the enantiomer of 6 (ent-6), as depicted in
Scheme 3. All that now separated these two compounds, 35
and 36, from the targeted diastereomeric structures of azaspi-
racid-1, 1a,b, was the casting of the final ring (E) and a few
accelerating effect would result in the desired compromise
between reaction rate acceleration and stereochemical integrity
preservation at C-25.14 Indeed, with some fine-tuning of
stoichiometry, we were able to achieve a 66% yield of coupling
product 30 without any observable epimerization. In this respect,
functional group adjustments.
The final stages of the drive toward the targeted azaspiracid-1
structures, 1a,b, are shown in Scheme 4. The secondary alcohol
in 35 was protected as a TES ether (TESOTf, 2,6-lutidine, 89%
yield) to allow selective oxidation at C-1. Thus, cleavage of
the C-1 acetate (K2CO3, MeOH, 85% yield), followed first by
a Swern oxidation [(COCl)2, DMSO; Et3N] and then by a
NaClO2 oxidation (81% overall yield for the two steps), led to
carboxylic acid 43 through intermediates 37, 39, and 41. The
TES group was then removed from 43 by treatment with TBAF
to afford acetoxy compound 45 (87% yield), a substrate thought
to be separated from the final target (1a) by only one step,
deacetylation. Indeed, on exposure of this compound (45) to
LiOH, the acetate was cleaved and an inseparable mixture18 of
it was interesting to note both the acceleration of the reaction
and the increase in yield as the amount of AsPh3 was decreased
from 80 to 10 mol %, matching the mol % of the palladium
catalyst [Pd(dba)3] (entries 3-6, Table 2).
Boasting all carbons needed for the azaspiracid-1 molecule,
intermediate 30 was then ready for further advancement, as
shown in Scheme 3. Because the relative stereochemistry
between the ABCDE and FGHI domains was not known at the
time, we went through the sequence with both isomers of the
FGHI domain (6 and ent-6). The next task was to forge ring G
and thus complete the FGHI domain of the molecule. Removal
of the TES group from the C-34 oxygen generated hydroxy enol
ether 4, whose iodoetherification reaction (NIS, NaHCO3)15
(14) Farina, V.; Krishman, B. J. Am. Chem. Soc. 1991, 113, 9585.
(15) (a) Adinolfi, M.; Parrilli, M.; Barone, G.; Laonigro, G.; Mangoni, L.
Tetrahedron Lett. 1976, 3661. (b) Haaima, G.; Weavers, R. T. Tetrahedron
Lett. 1988, 29, 1085.
(16) Carnevale, G.; Davini, E.; Iavarone, C.; Trogolo, C. J. Chem. Soc., Perkin.
Trans. 1: Org. Bioorg. Chem. 1990, 989.
(17) Miura, K.; Ichinose, Y.; Nozaki, K.; Fugami, K.; Oshima, K.; Utimoto, K.
Bull. Chem. Soc. Jpn. 1989, 62, 143.
9
J. AM. CHEM. SOC. VOL. 128, NO. 7, 2006 2263

A R T I C L E S
Nicolaou et al.
Scheme 4. Final Stages and Completion of the Total Synthesis of the Proposed Azaspiracid-1 Structures 1a and 1ba
a Reagents and conditions: (a) TESOTf (10 equiv), 2,6-lutidine (20 equiv), CH2Cl2, -78 f 0 °C, 10 min, 89% for 37 and 81% for 38; (b) K2CO3
(10 equiv), MeOH, 25 °C, 2 h, 85% for 39 and 87% for 40; (c) (COCl)2 (10 equiv), DMSO (20 equiv), CH2Cl2, -78 °C, 1 h; then Et3N (50 equiv),
-78 f -20 °C, 30 min; (d) NaClO2 (10 equiv), NaH2PO4 (10 equiv), 2-methyl-2-butene (excess), t-BuOH:H2O (4:1), 25 °C, 30 min, 81% for 41 and 79%
for 42 over two steps; (e) TBAF (5.0 equiv), THF, 25 °C, 2 h, 87% for 45, 92% for 46; (f) LiOH (10 equiv), MeOH:H2O (5:1), 25 °C, 16 h, 45% (for the
total inseparable mixture of at least two compounds represented by 1a) and 37% (for the total inseparable mixture of at least two compounds represented
by 1b).
Scheme 5. Inversion of C-20 Stereocentera
a Reagents and conditions: (a) Tf2O (10 equiv), py (20 equiv), 4-DMAP (0.2 equiv), CH2Cl2, 0 °C, 1.5 h, 68%; (b) KNO2 (5.0 equiv),
18-crown-6 (5.0 equiv), DMF, 25 °C, 24 h, 45%. Abbreviations: Tf, trifluoromethanesulfonyl; 4-DMAP, 4-(dimethylamino)pyridine; DMF, dimethylform-
amide.
products (at least two) was obtained, none of which matched
the natural product by TLC or 1H NMR spectroscopy. The same
sequence was then employed to advance the second diastereomer
36 to its final destination, 1b, through intermediates 38, 40, 42,
44, and 46, but again the obtained inseparable mixture (at least
two components) did not contain any compound matching the
natural substance by TLC or 1H NMR spectroscopy. Although
we had no definitive evidence to support a claim that precursors
45 and 46 led to structures 1a and 1b either as transient species
or as components of the obtained mixtures upon treatment with
LiOH, the fact that the natural substance was not present in
the mixture was clear, for neither its H NMR signals were
1
observed in the spectrum of the isolated mixture of com-
pounds nor did it show up in the TLC experiments that in-
cluded an authentic sample. The conclusion that the originally
proposed structures (i.e., 1a,b) were wrong was also supported
1
by the H NMR spectra of compounds 45 and 46 and the
(18) Separation could not be obtained either by preparative TLC (silica gel,
CH3Cl:CH3OH:H2O 20:3:1) or by HPLC according to the conditions of
Satake et al.2
9
2264 J. AM. CHEM. SOC. VOL. 128, NO. 7, 2006

Synthesis and Structure of Azaspiracid-1
A R T I C L E S
Scheme 6. Synthesis of the C-20 Epimeric Octacyclic Systems 54 and 55a
a Reagents and conditions: (a) Pd2dba3 (0.3 equiv), AsPh3 (0.3 equiv), LiCl (6.0 equiv), i-Pr2NEt (12.0 equiv); then 6 or ent-6 (0.03 M in THF, syringe
pump addition), NMP, 45 °C, 4 h, 58% for 48, 61% for 49; (b) HF‚py (excess), THF:pyridine (1:1), 0 f 25 °C, 2.5 h; (c) NIS (10 equiv), NaHCO3 (30
equiv), THF, 0 °C, 16 h, 68% for 52, 60% for 53 over two steps; (d) Et3B (1.0 M in hexanes, 3.0 equiv), n-Bu3SnH:toluene (1:2), 0 °C, 5 min, 79% for 54,
85% for 55.
intermediates leading up to them, which consistently exhibited
significant differences in the proton chemical shifts from the
ones reported for the natural substance, especially for H-6, H-7,
H-8, and H-9.
inversion of configuration at C-20,21 leading to 47 in 45% yield
(Scheme 5).
Schemes 6 and 7 summarize the chemistry used to arrive at
the azaspiracid C-20 epimers (2 and 3) following the previously
described chemistry (see Scheme 3). Unfortunately, as with 1a,b,
the last step (LiOH-induced cleavage of the C-25 acetate) proved
problematic in terms of yielding a single product.18 As before,
and as we expected, no component of the resulting mixtures
corresponded to natural azaspiracid-1 by TLC or 1H NMR
spectroscopic analysis. Furthermore, and as we brought it to an
end, this excursion, like the ones before it, left us with no clue
as to the true structure of the natural product.
3. Synthesis of Two of the C-20 Epimers of the Originally
Proposed Structures and a Second-Generation Late-Stage
Sequence. At this stage we decided to pursue the synthesis of
the C-20 epimers of structures 1a,b, despite our suspicions that
there was more to the structural discrepancy between these
1
structures and that of the natural product (based on H NMR
spectroscopic differences). We did this because it was the most
immediate option available to us at the time, and also in the
hope that, even if it did not tell the entire story, this expedition
may guide us, at least to some extent, in the right direction.
The inversion of the C-20 hydroxyl group proved nontrivial,
as numerous attempts to bring this about with Mitsunobu in-
version,19 stereoselective reduction20 of the C-20 carbonyl
compound, or SN2-type displacement employing substrates with
several leaving groups were met with failure. It was finally
found that triflate formation (Tf2O, py, 4-DMAP, 68% yield
from 29), followed by reaction of the intermediate triflate with
KNO2 in the presence of 18-crown-6, resulted in complete
Convinced of the fact that the originally proposed structural
assignments for azaspiracid-1 were in error, but troubled by our
failure to produce clean samples of these structures in the final
operation, we set out to modify the final stages of the synthesis
so as to avoid exposure of the precursors to LiOH in the final
step. Since these advanced intermediates proved robust to TBAF
conditions, we targeted the C-25 TBS ether as the penultimate
precursor to the targeted structures, and thus followed the
sequence shown in Scheme 8. The diacetate 66 was selectively
prepared from tetraol 28 by treatment with AcCl in the presence
(19) (a) Mitsunobu, O. Synthesis 1981, 1. (b) Hughes, D. L. Org. React. 1992,
42, 335.
(20) Corey, E. J.; Helal, C. J. Angew. Chem., Int. Ed. 1998, 37, 1987.
(21) Jain, R.; Kamau, M.; Wang, C.; Ippolito, R.; Wang, H.; Dulina, R.;
Anderson, J.; Gange, D.; Sofia, M. J. Bioorg. Med. Chem. Lett. 2003, 13,
2185.
9
J. AM. CHEM. SOC. VOL. 128, NO. 7, 2006 2265

A R T I C L E S
Nicolaou et al.
Scheme 7. Final Sequences of the Total Syntheses of C-20-epi-1a (2) and C-20-epi-1b (3)a
a Reagents and conditions: (a) TESOTf (10 equiv), 2,6-lutidine (20 equiv), CH2Cl2, -78 f 0 °C, 10 min; (b) K2CO3 (10 equiv), MeOH, 25 °C, 2 h; (c)
(COCl)2 (10 equiv), DMSO (20 equiv), CH2Cl2, -78 °C, 1 h; then Et3N (50 equiv), -78 f -20 °C, 30 min; (d) NaClO2 (10 equiv), NaH2PO4 (10 equiv),
2-methyl-2-butene (excess), t-BuOH:H2O (4:1), 25 °C, 30 min, 42% for 62 and 51% for 63 over four steps; (e) TBAF (5.0 equiv), THF, 25 °C, 2 h, 77%
for 64, 80% for 65; (f) LiOH (10 equiv), MeOH:H2O (5:1), 25 °C, 16 h, 45% (for the total inseparable mixture of at least two compounds represented by
2) and 38% (for the total inseparable mixture of at least two compounds represented by 3).
of 2,4,6-collidine in CH2Cl2 at -78 °C (89% yield) and
converted to the C-25 TBS ether 67 with TBSOTf and 2,6-
lutidine (82% yield). Oxidative cleavage of the dithiane moiety
from the latter compound [PhI(OCOCF3)2, 61% yield] then
allowed the generation of hydroxy ketone-allylic acetate 68,
which underwent smooth Stille coupling with stannane 6 as
before (Pd2dba3, LiCl, AsPh3, i-Pr2NEt, 63% yield) to afford
coupled product 69. The TES group was then selectively
removed from 69 by the action of TBAF (88% yield), and the
resulting dihydroxy compound was subjected to the iodoetheri-
fication conditions developed previously (NIS, NaHCO3),
affording iodoether 71 in 60% yield. Reductive removal of the
iodide residue from the latter comound (71) then led to
octacyclic intermediate 72 (55% yield), whose advancement to
carboxylic acid 76 proceeded smoothly as with the previous
series, and through intermediates 73 (TESOTf, 2,6-lutidine, 90%
yield), 74 (K2CO3, MeOH, 79% yield), and 76 [(COCl)2,
DMSO; Et3N, followed by NaClO2 oxidation, 80% yield over
the two steps]. The three silicon-based protecting groups
guarding the originally proposed structure 1a were then removed
by exposure to TBAF, but the product was, once again, an
inseparable mixture of at least two compounds,18 none of which
matched natural azaspiracid-1 by TLC or 1H NMR spectroscopy.
Changing the fluoride source for the final step (e.g., TASF, HF‚
py, HF‚Et3N) did not improve the outcome.
Conclusion
Having carried out the above-described investigations, we had
proven the mistaken structural identity originally ascribed to
azaspiracid-1, but we were nowhere near knowing the real
structure of this fascinating substance. Where was, or were, the
mistake(s)? That was the lingering question, one that could not
be answered through the information available to us at the time.
At this juncture, we concluded that wherever the error(s), a
drastically new approach had to be injected into our synthetic
program if we were to make headway toward elucidating the
azaspiracid-1 riddle. The new inspiration and assistance would
come from the natural substance itself, despite its extreme
scarcity, as we will describe in a subsequent publication.3
Acknowledgment. We thank Drs. D. H. Huang, G. Siuzdak,
and R. Chadha for NMR spectroscopic, mass spectrometric, and
X-ray crystallographic assistance, respectively. Financial support
for this work was provided by The Skaggs Institute for Chemical
Biology, the National Institutes of Health (USA), predoctoral
9
2266 J. AM. CHEM. SOC. VOL. 128, NO. 7, 2006

Synthesis and Structure of Azaspiracid-1
A R T I C L E S
Scheme 8. Second Generation Approach toward the Originally Proposed Structure 1aa
a Reagents and conditions: (a) AcCl (20 equiv), 2,4,6-collidine (40 equiv), CH2Cl2, -78 °C, 6 h, 89%; (b) TBSOTf (10 equiv), 2,6-lutidine (20 equiv),
CH2Cl2, -78 f 0 °C, 30 min, 82%; (c) PhI(OCOCF3)2 (2.0 equiv), MeCN:pH 7 buffer (4:1), 0 °C, 61%; (d) 68, Pd2dba3 (0.3 equiv), AsPh3 (0.3 equiv),
LiCl (6.0 equiv), i-Pr2NEt (12 equiv); then 4 (3.0 equiv, 0.03 M in THF, syringe pump addition), NMP, 40 °C, 1 h, 63%; (e) TBAF (1.0 M in THF, 1.2
equiv), THF, 0 °C, 20 min, 88%; (f) NIS (2.0 equiv), NaHCO3 (10 equiv), THF, 0 °C, 12 h, 60%; (g) Et3B (1.0 M in hexanes, 0.2 equiv), n-Bu3SnH:toluene
(1:2), 0 °C, 5 min, 55%; (h) TESOTf (10 equiv), 2,6-lutidine (20 equiv), CH2Cl2, -78 f 0 °C, 10 min, 90%; (i) K2CO3 (10 equiv), MeOH, 25 °C, 2 h, 79%;
(j) (COCl)2 (10 equiv), DMSO (20 equiv), CH2Cl2, -78 °C, 1 h; then Et3N (50 equiv), -78 f -20 °C, 30 min; (k) NaClO2 (10 equiv), NaH2PO4 (10 equiv),
2-methyl-2-butene (excess), t-BuOH:H2O (4:1), 25 °C, 30 min, 80% over two steps; (l) TBAF (5.0 equiv), THF, 25 °C, 15 h, 45% (for the total inseparable
mixture of at least two compounds represented by 1a). Abbreviation: TBS, tert-butyldimethylsilyl.
fellowships from Bristol-Myers Squibb, The Skaggs Institute
for Research, and The Scripps Research Institute Society of
Fellows (all to F. B.), the National Science Foundation (M. O.
F.) and Eli Lilly (Y. L.), postdoctoral fellowships from the Ernst
Schering AG Foundation (S. V.) and the Japanese Society for
the Advancement of Science (Y. M. A. Y.), and grants from
Abbott, Amgen, ArrayBiopharma, Boehringer-Ingelheim, Glaxo
Smith Kline, Hoffmann-LaRoche, DuPont, Merck, Novartis,
Pfizer, and Schering Plough.
Supporting Information Available: Experimental procedures
and compound characterization (CIF, PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
JA054748Z
9
J. AM. CHEM. SOC. VOL. 128, NO. 7, 2006 2267