Chemistry and biology of diazonamide A: First total synthesis and confirmation of the true structure

Article abstract of DOI:10.1021/ja040092i

With the addition of a tenth ring, the exchange of an oxygen atom for a nitrogen in the heart of the molecule, and a different terminal residue, the revised architecture for diazonamide A (1) provided an even more challenging molecular puzzle for chemical synthesis than its predecessor. In this article, we detail the first successful total synthesis of diazonamide A, an endeavor which not only verified its proper connectivities and established the stereochemistry of its previously unassignable C-37 chiral center, but which also was attended by the development of several new synthetic strategies and tactics.

Full text of DOI:10.1021/ja040092i

Published on Web 09/18/2004
Chemistry and Biology of Diazonamide A: First Total
Synthesis and Confirmation of the True Structure
K. C. Nicolaou,* David Y.-K. Chen, Xianhai Huang, Taotao Ling, Marco Bella, and
Scott A. Snyder
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 March 29, 2004; E-mail: kcn@scripps.edu
Abstract: With the addition of a tenth ring, the exchange of an oxygen atom for a nitrogen in the heart of
the molecule, and a different terminal residue, the revised architecture for diazonamide A (1) provided an
even more challenging molecular puzzle for chemical synthesis than its predecessor. In this article, we
detail the first successful total synthesis of diazonamide A, an endeavor which not only verified its proper
connectivities and established the stereochemistry of its previously unassignable C-37 chiral center, but
which also was attended by the development of several new synthetic strategies and tactics.
Introduction
Although structural revisions of natural products are quite
common, few of these reassignments occur with as many
profound changes as the ones applied to diazonamide A (1,
Figure 1). Indeed, in addition to the complete alteration of the
terminal amino acid attached to the C-2 amine and the
incorporation of a tenth ring in the form of an aminal (ring H),
what was previously thought to be an oxygen atom at its
epicenter was exchanged for a nitrogen. As such, any synthetic
plan that had been developed for diazonamide A’s originally
proposed structure (see previous articles)1 would need to be
significantly retooled to gain access to building blocks ap-
propriately functionalized for the new target. Once these
fragments were built, however, the total synthesis of 1 was not
necessarily near or guaranteed. For example, as experience
garnered during campaigns toward the original structure had
consistently demonstrated, while it was relatively easy to
construct fragments bearing the disparate elements of the
diazonamide structure, it was quite difficult to merge them into
structures containing its highly strained and rigid 12-membered
macrocyclic systems.2
Figure 1. The revised structural assignment of diazonamide A (1). Note:
the C-37 stereochemistry is unassigned.
our chances for success. Our first approach was patterned after
the chemistry which had been developed to access the originally
proposed structure of diazonamide A as described in the
previous two articles in this series.1 According to this plan, we
would attempt to form the heterocyclic core first through a
heteropinacol macrocyclization, and then build the second
macrocycle of 1 through macrolactamization. The second plan
of attack, which is the subject of this article, sought to reach
diazonamide A (1) via the alternate order of macrocycle
formation. Our hope was that at least one of these approaches
would prove capable of delivering diazonamide A (1) to verify
its structure and, if both plans worked in part to produce one of
the 12-membered rings, then we would be able to generate
simplified analogues of both domains for the purposes of
biological investigations.
Faced with this conundrum, we elected to confront the
structural complexity posed by the “new” diazonamide A (1)
from not one, but two, completely different directions to enhance
(1) (a) Nicolaou, K. C.; Snyder, S. A.; Huang, X.; Simonsen, K. B.; Koumbis,
A. E.; Bigot, A. J. Am. Chem. Soc. 2004, 126, 10162-10173. (b) Nicolaou,
K. C.; Snyder, S. A.; Giuseppone, N.; Huang, X.; Bella, M.; Reddy, M.
V.; Bheema Rao, P.; Koumbis, A. E.; Giannakakou, P.; O’Brate, A. J.
Am. Chem. Soc. 2004, 126, 10174-10182.
(2) In truth, only a few of the creative strategies developed toward the original
structure managed to identify a sequence that could generate either of these
12-membered systems in any measurable yield. See, for example: (a)
Nicolaou, K. C.; Snyder, S. A.; Simonsen, K. B.; Koumbis, A. E. Angew.
Chem., Int. Ed. 2000, 39, 3473-3478. (b) Nicolaou, K. C.; Huang, X.;
Giuseppone, N.; Bheema Rao, P.; Bella, M.; Reddy, M. V.; Snyder, S. A.
Angew. Chem., Int. Ed. 2001, 40, 4705-4709. (c) Vedejs, E.; Zajac, M.
A. Org. Lett. 2001, 3, 2451-2454. (d) Reference 3 in this article.
Results and Discussion
Retrosynthetic Analysis. Our new retrosynthetic plan for
diazonamide A (1) is shown in Scheme 1, starting with
essentially the same modifications that we had applied in our
analysis of the originally proposed structure to prepare the
ground for retrons seeking to open its macrocyclic systems.
These were the removal of the C-2 hydroxyisovaleric acid side
9
12888
J. AM. CHEM. SOC. 2004, 126, 12888-12896
10.1021/ja040092i CCC: $27.50 © 2004 American Chemical Society

Diazonamide A: First Total Synthesis
A R T I C L E S
Scheme 1. General Retrosynthetic Analysis of Diazonamide A (1) on the Basis of Constructing the Heterocycle-Based Macrocycle Last
chain (2), since its lone chiral center (C-37) could not be
assigned,3 and the excision of the two aryl chlorines and the
FH aminal ring on the basis of considerations of their stability
to diverse reaction conditions. In the forward direction, we
expected that these latter two motifs could be generated in either
order, with the chloro groups arising through a site-selective
electrophilic aromatic substitution reaction and the aminal ring
resulting from the addition of the C-7 phenol onto an iminium
species derived from the F-ring lactam in 3.
Having reached an appropriate subgoal structure to tackle the
question of macrocycle formation, we then unfurled the
heterocycle-based ring system at its C16-C18 biaryl linkage
through a Witkop-type photocyclization transform.4 This same
disconnection had been used to great success by the Harran
group in their synthesis of the originally proposed structure.3,5
Accordingly, the chances that 4 could be converted into 3 during
the actual synthesis were relatively high, although the presence
of new functionality within the E-ring of 4 meant that some
modification of the reaction conditions would likely be needed
to promote the single-electron-transfer step in this case. Equally
important, from the standpoint of design, the macrocyclization
should occur atropselectively as governed by the chirality of
the C-10 quaternary center and the likely strong driving force
to orient the B- and E-rings in a stacked arrangement through
π-π interactions (an orientation that only the correct atrop-
isomer can accomplish).
Accepting these key alterations, the heterocyclic chain of this
new target (4) was then simplified to 7 by unraveling its B-ring
oxazole to a keto amide precursor (5) and breaking apart the
(3) (a) Li, J.; Jeong, S.; Esser, L.; Harran, P. G. Angew. Chem., Int. Ed. 2001,
40, 4765-4770. (b) Li, J.; Burgett, A. W. G.; Esser, L.; Amezcua, C.;
Harran, P. G. Angew. Chem., Int. Ed. 2001, 40, 4770-4773.
(4) (a) Yonemitsu, O.; Cerutti, P.; Witkop, B. J. Am. Chem. Soc. 1966, 88,
3941-3945. (b) Theuns, H. G.; Lenting, H. B. M.; Salemink, C. A.; Tanaka,
H.; Shibata, M. S.; Ito, K.; Lousberg, R. J. J. Heterocycles 1984, 22, 2007-
2011. (c) Mascal, M.; Moody, C. J. J. Chem. Soc., Chem. Commun. 1988,
589-590.
(5) For the third successful total synthesis of diazonamide A, one which also
utilized a Witkop reaction to great success, see Burgett, A. W. G.; Li, Q.;
Wei, Q.; Harran, P. G. Angew. Chem., Int. Ed. 2003, 42, 4961-4966.
9
J. AM. CHEM. SOC. VOL. 126, NO. 40, 2004 12889

A R T I C L E S
Nicolaou et al.
Scheme 2. Precedent for the Reaction of Aromatics with Isatin
(14) under Acidic Conditions
heterocyclic chemistry, the successful translation of these studies
to the diazonamide context was not clear, especially in light of
the fact that an oxazole ring (ring A) and a phenol (ring G)
have far different reactivities.
Upon further consideration, we wondered if success could
be achieved by forming a compound such as 16 through a
different reaction, and then introducing the second aromatic
moiety through the acid-catalyzed electrophilic aromatic sub-
stitution process that Baeyer and Olah had used so productively
in the past. As shown in Scheme 1, this idea was expressed
through the initial nucleophilic addition of a dianion derived
from oxazole 12 onto the more reactive C-3 carbonyl group of
isatin 13, a process that would lead to the requisite test
compound (11) bearing a tertiary hydroxyl group. Subsequent
exposure of this product (11) to acid in the presence of the
electron-rich L-tyrosine-derived building block 9 was then
expected to give rise to 8 by way of 10. Of course, while this
strategy would afford an expedient entry into the diazonamide
skeleton if it worked, it could never lead to the stereocontrolled
synthesis of the C-10 center in the absence of additional tricks.
However, because we found it challenging to envision a more
expeditious synthesis of such a fragment, the loss of a portion
of material because of shortcomings imposed by its conciseness
was anticipated to provide a level of material throughput that
could match, or even surpass, a longer route featuring stereo-
control.8 Thus, the formidable architecture of diazonamide A
(1) had been reduced to five simple building blocks (2, 6, 9,
12, and 13) in what amounted to a convergent synthetic plan.9
Preliminary Model Studies. We began our evaluation of
this plan’s feasibility with several model studies directed toward
the construction of the C-10 quaternary center of diazonamide
A with all of its peripheral structural motifs. Pleasingly, as
indicated in Scheme 3, such travail would ultimately bear fruit
in the form of a successful construction of 27.
The first key operation was merging Cbz-protected L-tyrosine
methyl ester10 and isatin derivative 22 [the product of chemose-
lective Grignard addition of phenylmagnesium bromide onto
7-bromoisatin (21)11] through electrophilic aromatic substitution.
At first, forming this key carbon-carbon bond proved elusive,
as treatment of 1.00 equivalents of 22 and 1.05 equivalents of
23 with a variety of acids in catalytic quantities led exclusively
to decomposition (Table 1, entries 1-7). Thankfully, though,
when the amount of acid was increased to stoichiometric (entry
newly unveiled amide linkage to excise the CD-indole portion
as amine 6. With the formidable heterocycle-based macrocycle
of 1 effectively stripped away, we now needed to dissect the
remaining 12-membered ring of diazonamide A (1). As indi-
cated, we unlocked this ring at its seemingly most logical site,
its central amide, to provide an amino acid precursor (8)
modified by the introduction of a few protecting groups.
However, while this choice was a relatively obvious one, the
operations that could, in turn, convert this new target (8) into
simple building blocks were far from self-evident since that task
entailed the detachment of three disparate aromatic systems from
a quaternary carbon. Indeed, literature precedent for how to form
such systems was relatively sparse.
For example, at the end of the 19th century, Baeyer and
Lazarus were able to attach two identical aromatic systems to
a molecule of isatin (14, Scheme 2) through iterative electro-
philic aromatic substitution reactions proceeding by way of
intermediates 15-18.6 Over one hundred years later, a report
from Olah and co-workers described how the same process
could be employed to attach two different aromatics to isatin
(14), albeit in random fashion, by using two aromatic reactants
simultaneously to obtain a statistical mixture of all three possible
products favoring 20.7 While both were important advances in
(8) Several such outcomes are known. For instance, in their elegant studies
toward the natural product meso-chimonanthine, the Overman group
developed a 13-step asymmetric synthesis of this compound which
proceeded in an overall yield of ∼30%: (a) Overman, L. E.; Larrow, J. F.;
Stearns, B. A.; Vance, J. M. Angew. Chem., Int. Ed. 2000, 39, 213-215.
(b) Overman, L. E.; Paone, D. V.; Stearns, B. A. J. Am. Chem. Soc. 1999,
121, 7702-7703. Later, another group developed a biomimetic, three-step
approach to the same target. Although it lacked any stereocontrol, it afforded
the same product in 30% yield following separation from the undesired
stereoisomers: Ishikawa, H.; Takayama, H.; Aimi, N. Tetrahedron Lett.
2002, 43, 5637-5639.
(9) For the preliminary communication of this total synthesis, see Nicolaou,
K. C.; Bella, M.; Chen, D. Y.-K.; Huang, X.; Ling, T.; Snyder, S. A. Angew.
Chem., Int. Ed. 2002, 41, 3495-3499.
(10) The synthesis of this compound was achieved by Cbz-protection of
L-tyrosine methyl ester hydrochloride using benzyl chloroformate and
Na2CO3 as described in ref 2a.
(11) 7-Bromoisatin was prepared via the standard Sandmeyer procedure, namely,
reacting 2-bromoaniline with chloral hydrate and HONH2‚HCl, and then
cyclizing the resultant product with H2SO4: (a) Lisowski, V.; Robba, M.;
Rault, S. J. Org. Chem. 2000, 65, 4193-4194. (b) Ma, D.; Wu, Q.
Tetrahedron Lett. 2000, 41, 9089-9093. (c) Tokunaga, T.; Hume, W. E.;
Umezome, T.; Okazaki, K.; Ueki, Y.; Kumagai, K.; Hourai, S.; Nagamine,
J.; Seki, H.; Taiji, M.; Noguchi, H.; Nagata, R. J. Med. Chem. 2001, 44,
4641-4649.
(6) Baeyer, A.; Lazarus, M. J. Chem. Ber. 1885, 18, 2637-2643.
(7) Klumpp, D. A.; Yeung, K. Y.; Prakash, G. K. S.; Olah, G. A. J. Org.
Chem. 1998, 4481-4484.
9
12890 J. AM. CHEM. SOC. VOL. 126, NO. 40, 2004

Diazonamide A: First Total Synthesis
A R T I C L E S
Scheme 3. Model Studies Exploring the Generation of the C-10
Quaternary Center of the Revised Structure of Diazonamide A (1)a
Table 2. Attempts to Form the FH Aminal Ring System of a
Diazonamide A Model System through Reductive Closurea
entry
conditions
yield (%)
1
2
3
4
5
6
7
DIBAL-H,b THF, 25 °C
Red-Al, THF, 25 °C
LiBEt3H, THF, 25 °C
BH3‚SMe2, THF, 25 °C
NaBH4, THF, 25 °C
BH3‚THF, THF, 25 °C
23
31
32
40
37
26
18
TMSOTf (1.0 equiv), Et3SiH (20 equiv), THF, 0f25 °C
a Compounds 24 and 27 are mixtures of C10 epimers (ca. 1:1). b DIBAL-
H ) diisobutylaluminum hydride.
of 24 to virtually any of the common reducing agents (see Table
2) led to the formation of the bicyclic aminal system 27,
presumably by way of intermediates 25 and 26 [Compounds
24-27 are mixtures of C-10 epimers (ca. 1:1)].12 Although the
best yield that we obtained in this final step was only 40%
(unoptimized), we were hopeful that when it came time to close
the same ring system with fully elaborated intermediates, the
restricted rotational freedom of the G-ring phenol as part of a
completed left-hand macrocyclic unit would drive the ring
closure with higher efficiency than that observed here. In any
case, we knew that we could form the C-10 quaternary center
of diazonamide A (1) through this approach, and we had
concurrently identified a method that could be applied in the
late stages of the synthesis to form its aminal ring.
a Reagents and conditions: (a) PhMgBr (2.5 equiv), THF, 0 °C, 96%;
(b) 23 (1.05 equiv), TfOH (5.0 equiv), CH2Cl2, 0 °C, 20 min; then Ac2O
(10 equiv), CH2Cl2, 25 °C, 20 min, 84%; (c) BH3‚SMe2 (4.0 equiv), THF,
25 °C, 2 h, 40%. TfOH ) trifluoromethanesulfonic acid. Compounds 24-
27 are mixtures of C10 epimers (ca. 1:1).
Table 1. Attempts to Form the Triaryl-Substituted Quaternary
Center of a Diazonamide A (1) Model System through Electrophilic
Aromatic Substitutiona
Synthesis of Building Blocks. With these studies complete,
we sought to apply our designed sequence to building blocks
appropriately decorated to access diazonamide A (1), a task
which required supplies of 2-hydroxyisovaleric acid (2), keto
amide 6, oxazole 12, and MOM-protected 7-bromoisatin (13,
see Scheme 1). Of these, only the last three demanded any route
development since the first was commercially available.
The construction of oxazole fragment 12 is shown in Scheme
4, starting with the merger of Boc-protected L-valine (28) and
DL-serine methyl ester (29) into 30 in 92% yield through a
standard peptide-coupling reaction as conducted by EDC and
HOBt. Although this reaction was a routine one, it served to
provide a product (30) bearing all the structural motifs required
to cast the target’s oxazole ring. Indeed, following cyclization
to oxazoline 31 through the action of Burgess reagent13 in
entry
conditions
yield (%)
1
2
3
4
5
6
7
8
Yb(OTf)3 (cat.), CH2Cl2, 25 °C
Sc(OTf)3 (cat.), CH2Cl2, 25 °C
pTsOH (1.0 equiv), CH2Cl2, 25 °C
pTsOH (1.0 equiv), toluene, 25 °C
CSAb (1.0 equiv), CH2Cl2, 25 °C
PPTSc (1.0 equiv), CH2Cl2, 25 °C
TfOH (0.2 equiv), CH2Cl2, 25 °C
MeSO3H/CH2Cl2 (1:5), 0 °C
decomposition
decomposition
decomposition
decomposition
decomposition
decomposition
decomposition
47
9
10
TfOH (2 equiv), CH2Cl2, 25 °C
TfOH/CH2Cl2 (1:5), 25 °C
63
84
(12) For examples of related iminium closures, see (a) Shishido, K.; Shitara,
E.; Komatsu, H.; Hiroya, K.; Fukumoto, K.; Kametani, T. J. Org. Chem.
1986, 51, 3007-3011. (b) Shishido, K.; Hiroya, K.; Komatsu, H.;
Fukumoto, K.; Kametani, T. J. Chem. Soc., Perkin Trans. 1 1987, 2491-
2495. (c) Ezquerra, J.; Pedregal, C.; Yruretagoyena, B.; Rubio, A.; Carren˜o,
M. C.; Escribano, A.; Ruano, J. L. G. J. Org. Chem. 1995, 60, 2925-
2930. (d) Collado, I.; Ezquerra, J.; Mateo, A. I.; Rubio, A. J. Org. Chem.
1998, 63, 1995-2001. (e) Downham, R.; Ng, F. W.; Overman, L. E. J.
Org. Chem. 1998, 63, 8096-8097.
(13) For reviews on the chemistry of the Burgess reagent, see (a) Taibe, P.;
Mobashery, S. In Encyclopedia of Reagents for Organic Synthesis, Vol. 5;
Paquette, L. A., Ed.; John Wiley & Sons: Chichester, U.K., 1995; pp 3345-
3347. (b) Burckhardt, S. Synlett 2000, 559. For the initial report of this
reagent, see (c) Atkins, G. M.; Burgess, E. M. J. Am. Chem. Soc. 1968,
90, 4744-4745. (d) Atkins, G. M.; Burgess, E. M. J. Am. Chem. Soc. 1972,
94, 6135-6141. (e) Burgess, E. M.; Penton, H. R.; Taylor, E. A. J. Org.
Chem. 1973, 38, 26-31.
a Compound 24 is a mixture of C10 epimers (ca. 1:1). b CSA )
10-camphorsulfonic acid. c PPTS ) pyridinium p-toluenesulfonate.
8) or gross excess levels (entries 9 and 10), triaryl intermediate
24 could be obtained in yields as high as 84% following
acetylation of the free amine liberated during the event. We
hypothesize that the amount of acid required reflects a need to
fully ionize the indole-based starting material in a relatively
short period of time to get it to participate in the desired
pathway. With this critical stage reached, subsequent exposure
9
J. AM. CHEM. SOC. VOL. 126, NO. 40, 2004 12891

A R T I C L E S
Nicolaou et al.
Scheme 4. Synthesis of Oxazole Building Block 12a
Scheme 5. Synthesis of Building Blocks 6 and 13a
a Reagents and conditions: (a) 28 (1.0 equiv), 29 (1.0 equiv), EDC (1.05
equiv), HOBt (1.05 equiv), THF, 25 °C, 5 min; then Et3N (2.0 equiv), 25
°C, 2 h, 92%; (b) Burgess reagent (1.5 equiv), THF, 65 °C, 4 h, 80%; (c)
BrCCl3 (1.1 equiv), DBU (1.1 equiv), CH2Cl2, 0f25 °C, 3 h, 90%. EDC
) 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide, HOBt ) 1-hydroxy-
benzotriazole, DBU ) 1,8-diazabicyclo[5.4.0]undec-7-ene. Compound 31
is a mixture of diastereomers (ca. 1:1).
a Reagents and conditions: (a) (Boc)2O (1.1 equiv), aq NaHCO3/dioxane
(1:2), 25 °C, 2 h; (b) DDQ (2.0 equiv), THF/H2O (9:1), 0 °C, 2 h; (c) TFA,
25 °C, 10 min, 79% over three steps; (d) vinylmagnesium bromide (3.0
equiv), THF, -45 °C, 45 min, 63%; (e) KOH (2.5 equiv), DMF, 0 °C, 5
min; I2 (1.05 equiv), DMF, 0 °C, 15 min; (f) LiHMDS (1.0 M in THF, 1.2
equiv), THF, -78 °C, 15 min; MOMCl (1.4 equiv), -78f25 °C, 2 h; (g)
RuCl3‚3H2O (0.1 equiv), NaIO4 (3.0 equiv), CH3CN/THF/H2O (8.3:1.0:
1.7), 0 °C, 30 min, 68% over three steps. DDQ ) 2,3-dichloro-5,6-dicyano-
1,4-benzoquinone, TFA ) trifluoroacetic acid, DMF ) N,N-dimethylfor-
mamide, LiHMDS ) lithium bis(trimethylsilyl)amide, MOM ) methoxy
methyl.
refluxing THF over the course of 4 h, subsequent aromatization
as effected by BrCCl3 and DBU in CH2Cl2 at 25 °C delivered
the requisite fragment (12) in 72% overall yield.14 The targeted
fragment could also be obtained in the same number of steps
through Robinson-Gabriel cyclodehydration of the aldehyde
congener of 30, but in significantly reduced yields (∼30% over
two steps).15
fragments were smoothly converted into 11 (Scheme 6) in 73%
yield by treating a THF solution of that new oxazole (36) with
2.0 equiv of n-BuLi at -78 °C for 20 min, and then adding 13
to the reaction flask. With this operation unveiling a tertiary
hydroxyl group at the carbon corresponding to the C-10 position
of the target molecule, this new intermediate (11) was perfectly
outfitted to attempt the incorporation of the final aromatic ring
of the AG system in the next step. After extensive optimization
of the conditions originally identified for this task in our model
studies (vide supra), this goal was accomplished in 47% yield
by refluxing a solution containing 9 (4.0 equiv), 11 (1.0 equiv),
and p-TsOH (4.0 equiv) in 1,2-dichloroethane for 25 min.
Although the material return for this reaction was modest, this
shortcoming was not a measure of the efficiency of the reaction
itself but rather reflected the difficulty in isolating the final
product because of the presence of the free amine (38) which
had been unveiled unintentionally. Unable to improve this
outcome any further, we pressed forward and reprotected this
site in the next operation, now as a tert-butyl carbamate, to
provide 8 and 39 as a mixture of chromatographically separable
diastereomers in a combined yield of 76%. However, despite
our ability to obtain both C-10 epimers of this advanced
intermediate in stereochemically pure form following this
operation, we could not definitively assign their C-10 configura-
tions through any form of noncrystallographic analysis. Con-
sequently, both 8 and 39 were processed separately through the
ensuing steps of the sequence, hoping that at some stage their
physical data, or a crystal structure, would indicate which one
With one fragment in hand, we then sought to prepare the
remaining two. As shown in Scheme 5, we were able to access
indole-fragment 6 by selectively protecting the free primary
amine of tryptamine (32), installing a ketone carbonyl through
the action of DDQ in aqueous THF,16 and then exposing the
resultant product to TFA to concomitantly cleave the now-
extraneous Boc protecting group and deliver the product as its
stable TFA salt. MOM-protected 7-bromoisatin (13) required
almost as few steps to prepare as 6, ultimately arising after four
operations from 1-bromo-2-nitrobenzene, by way of 7-bromoin-
dole (34), in 43% yield. Although this fragment (13) could also
be prepared in two steps through MOM-protection of 7-bro-
moisatin (21, Scheme 3) formed by the more classical Sand-
meyer procedure,11 an inability to consistently secure large
amounts of commercial chloral hydrate on the basis of its recent
classification as a regulated substance led to the development
of the alternate approach shown here.
4. Synthesis of the First Macrocyclic Unit. Having ac-
complished this preparative work, attempts to merge the
synthesized pieces into the diazonamide skeleton began in
earnest, starting with the operations needed to effect the
generation of the C-10 center. At first, the initial task required
to achieve this goal, the addition of oxazole 12 to isatin 13,
met with failure because of an inability to generate a dianion
from the oxazole fragment without touching its ester. However,
when that motif was converted into a protected primary alcohol
(LiBH4, THF; then LiHMDS, TBAI, BnBr, THF),17 these
(14) Williams, D. R.; Lowder, P. D.; Gu, Y.-G.; Brooks, D. A. Tetrahedron
Lett. 1997, 38, 331-334.
(17) The use of LiHDMS was crucial in obtaining the desired benzylated product
(36) in high yield. Other bases such as NaH or LDA led to significant
amounts of bis-benzylated material by enabling engagement of the Boc-
protected amine. The reason for this strange difference in behavior,
especially as relates to the two lithium bases, is currently unknown.
(15) For the synthesis of a related fragment, see Downing, S. V.; Aguilar, E.;
Meyers, A. I. J. Org. Chem. 1999, 64, 826-831.
(16) Oikawa, Y.; Toshioka, T.; Mohri, K.; Yonemitsu, O. Heterocycles 1979,
12, 1457-1462.
9
12892 J. AM. CHEM. SOC. VOL. 126, NO. 40, 2004

Diazonamide A: First Total Synthesis
A R T I C L E S
Scheme 7. Completion of the Peptide-Based Macrocycle 7a
Scheme 6. Synthesis of Advanced Intermediate 8a
a Reagents and conditions: (a) MOMCl (20 equiv), K2CO3 (40 equiv),
acetone, 0f25 °C, 1 h, 85%; (b) LiOH (20 equiv), MeOH/THF/H2O (2:
10:1), reflux, 20 min, 96%; (c) TFA (neat), 10 min, 98%; (d) HATU (2.2
equiv), collidine (6.6 equiv), DMF/CH2Cl2 (1:2, 3.0 × 10-4 M), 25 °C, 12
h, 36%; (e) BCl3 (1.0 M in hexanes, 20 equiv), CH2Cl2, -78 °C, 2 h; then
THF, saturated aq NaHCO3, 10% aq NaOH, 25 °C, 1 h, 65%. HATU )
2-(1H-9-azabenzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophos-
phate.
a Reagents and conditions: (a) LiBH4 (2.0 M in THF, 3.0 equiv), EtOH
(10 equiv), Et2O/THF (2:1), 0f25 °C, 3 h, 83%; (b) NaHMDS (1.0 M in
THF, 2.0 equiv), THF, -78 °C, 10 min, then TBAI (0.1 equiv), BnBr (10
equiv), -78f25 °C, 24 h, 72%; (c) nBuLi (2.0 equiv), THF, -78 °C, 20
min, then 13 (1.05 equiv), 5 min, 73%; (d) 9 (4.0 equiv), pTsOH (4.0 equiv),
ClCH2CH2Cl, 83 °C, 25 min, 47%; (e) (Boc)2O (1.1 equiv), aq NaHCO3/
dioxane (1:2), 25 °C, 2 h, 65%. TBAI ) tetra-n-butylammonium iodide,
Bn ) benzyl, pTsOH ) p-toluenesulfonic acid, Boc ) t-butoxycarbonyl.
not especially surprising, the result that did confound us was
the observation that only one of the two C-10 epimers of 41
underwent macrocyclization successfully, with the other leading
primarily to dimerized material irrespective of reaction con-
centration.19 Thus, the troubling question raised by this finding
was whether it was the correct diastereomer that had cyclized
as drawn in Scheme 7, because if the wrong C-10 epimer of 42
had been formed, then we would never be able to access 1
through this strategy. Some positive evidence that the correct
epimer had indeed cyclized was provided once the MOM and
benzyl ethers attached to our mixture of products (42 and 43)
possessed stereochemistry corresponding to that of (-)-
diazonamide A (1).
The next objective of the sequence was to construct the only
major structural element still missing from the AG system of
1, namely, the amide bond that would complete its 12-membered
ring. Thus, to set the stage for the macrolactamization reaction
that would hopefully accomplish that goal, both 8 and 39 were
converted into amino acid 41 (see Scheme 7, shown as a single
sequence from 8) in near quantitative yield via initial protection
of their oxindole nitrogen and G-ring phenol as MOM ethers
(MOMCl, K2CO3, acetone), followed by ester hydrolysis and
Boc removal. Probes to form the final linkage of the AG
macrocycle then began. Despite the difficulty encountered in
forming this ring system in our studies to synthesize the
originally proposed structure for diazonamide A,1a no such
troubles were encountered here in the absence of a formed
heterocyclic core. Indeed, several conditions rose to the occa-
sion, with the use of HATU and 2,4,6-collidine in a 1:2 mixture
(18) For other total syntheses which benefited from this reagent combination,
see (a) Nicolaou, K. C.; Koumbis, A. E.; Takayanagi, M.; Natarajan, S.;
Jain, N. F.; Bando, T.; Li, H.; Hughes, R. Chem. Eur. J. 1999, 5, 2622-
2647. (b) Hu, T.; Panek, J. S. J. Org. Chem. 1999, 64, 3000-3001. (c)
Evans, D. A.; Wood, M. R.; Trotter, B. W.; Richardson, T. I.; Barrow, J.
C.; Katz, J. L. Angew. Chem., Int. Ed. 1998, 37, 2700-2704.
(19) Other examples of such selective cyclizations as dictated by the stereo-
chemistry of the intervening chain exist. See, for example, one of the classic
total syntheses of erythronolide: Woodward, R. B.; Logusch, E.; Nambiar,
K. P.; Sakan, K.; Ward, D. E.; Au-Yueng, B.-W.; Balaram, P.; Browne, L.
J.; Card, P. J.; Chen, C. H.; Cheˆnevert, R. B.; Fliri, A.; Frobel, K.; Gais,
H.-J.; Garratt, D. G.; Hayakawa, K.; Heggi, W.; Hesson, D. P.; Hoppe, D.;
Hoppe, I.; Hyatt, J. A.; Ikeda, D.; Jacobi, P. A.; Kim, K. S.; Kobuke, Y.;
Kojima, K.; Krowicki, K.; Lee, V. J.; Leutert, T.; Malchenko, S.; Martens,
J.; Matthews, R. S.; Ong, B. S.; Press: J. B.; Rajan Babu, T. V.; Rosseau,
G.; Sauter, H. M.; Suzuki, M.; Tatsuta, K.; Tolbert, L. M.; Truesdale, E.
A.; Uchida, I.; Ueda, Y.; Uychara, T.; Vasella, A. T.; Vlauchick, W. C.;
Wade, P. A.; Williams, R. M.; Wong, N.-C. J. Am. Chem. Soc. 1981, 103,
3213-3215.
of DMF and CH2Cl2 at a final concentration of 3.0 × 10-4
M
constituting the best protocol by delivering 42 along with a small
amount of phenol 43 in 36% combined yield.18 In the absence
of high dilution, we observed significant amounts of dimeric
and higher oligomeric materials. Although this outcome was
9
J. AM. CHEM. SOC. VOL. 126, NO. 40, 2004 12893

A R T I C L E S
Nicolaou et al.
Scheme 8. An Alternate Approach to Form the C-10 Quaternary
Center and the Peptide-Based Macrocycle Concomitantly Met with
Failurea
Table 3. Optimization of the Photochemical Conditions for Witkop
Closure
entry
conditionsa
yield (%)
1
2
3
4
5
6
7
8
quartz filter, Pyrex vessel, 200 W lamp, 30 min
quartz filter, Pyrex vessel, 200 W lamp, 12 h
borosilicate filter, Pyrex vessel, 200 W lamp, 30 min
sodium filter, Pyrex vessel, 200 W lamp, 30 min
quartz filter, Pyrex vessel, 450 W lamp, 30 min
quartz filter, Pyrex vessel, 450 W lamp, 12 h
sodium filter, Pyrex vessel, 450 W lamp, 10 min
sodium filter, quartz vessel, 450 W lamp, 12 min
5-10
dec
trace
5
15
dec
20
33
a All reactions were carried out in a 3:1 mixture of MeCN/H2O at 25 °C
in the presence of 3.0 equiv of epichlorohydrin and 2.0 equiv of LiOAc.
and D-rings, through an initial peptide coupling with 6 to afford
keto amide 27, followed by a Robinson-Gabriel cyclodehy-
dration using pyridine-buffered POCl3 at 25 °C. Interestingly,
other oxazole forming protocols, such as Cl3CCCl3, Ph3P, and
Et3N21 or the Burgess reagent,13 also accomplished the formation
of 4 from 5, but none afforded a yield commensurate to that
obtained with the above conditions which were developed as
part of our campaign toward the original structure of diazon-
amide A.1a,2
With the synthesis of compound 4 achieved in 34% overall
yield from 48, we now had in hand a substrate which could be
subjected to Witkop-type photocyclization conditions. Pleas-
ingly, little alteration of the protocol developed by the Harran
group3 to accomplish this ring closure in their synthesis of the
originally proposed structure of diazonamide A was required
here, with excitation at 200 nm in the presence of excess
epichlorohydrin (an acid scavenger) and LiOAc leading to the
assembly of 3 in a reproducible yield of 33%. Key to obtaining
this yield was the use of a quartz reaction vessel, a 450 W lamp,
and a short reaction time (15 min), as various alternatives (see
Table 3) either led to no product whatsoever or to a decreased
yield of 3. Importantly, though, when the reaction proceeded it
did so with complete atropselectivity for both the C16-C18
and C24-C26 biaryl axes. Equally significant, most of the
remaining material balance constituted recovered starting
material that could be recycled to enhance the overall throughput
of this process.22 As an additional note, we were also able to
convert 4 into 3 under strict radical conditions by treating 4
with a slight excess of Ph3SnH (2 equiv) and catalytic amounts
of AIBN in refluxing benzene. This alternate macrocyclization
proceeded in 10% yield.23
a Reagents and conditions: (a) TFA, 25 °C, 10 min; (b) 45 (1.5 equiv),
EDC (1.5 equiv), HOBt (1.5 equiv), 25 °C; then Et3N, 0 °C; then 25 °C,
12 h, 63% over two steps.
were removed through the action of BCl3 in CH2Cl2 at -78
1
°C, as the H NMR data of the resultant compound (7) was
remarkably similar to the natural product. While not definitive
proof, it was enough to induce us to continue our studies and
attempt to construct the second macrocyclic unit.
Before describing the next endeavors, it is important to
recognize that though the presence of the MOM groups
throughout the sequence shown in Scheme 7 may appear
unnecessary, they served to improve the yield of the ring closure
leading to 42 by facilitating this product’s isolation. In their
absence, we could never obtain more than a 15% yield of the
macrolactamization product. In addition, we also attempted to
execute several other strategies for macrocycle formation using
elements of this same sequence from intermediate 11, such as
ring closure through an acid-catalyzed intramolecular electro-
philic aromatic substitution of compound 46 (see Scheme 8).
Unfortunately, these approaches did not lead to the formation
of macrocycle 43 despite screening numerous activating can-
didates of both Brønsted and Lewis acidic types.20
5. Completion of the Synthesis. With a route to 7 and at
least some faith in its proper stereochemistry, work was now
directed toward the construction of the heterocyclic core. As
drawn in Scheme 9, these efforts began by selectively protecting
its H-ring phenol as a Boc carbonate using (Boc)2O in a solvent
mixture of aqueous NaHCO3 and 1,4-dioxane (1:2) over the
course of 24 h. With this operation leading to 47 in 91% yield
on the basis of the recovery of a small amount of starting
material (7), the remaining free hydroxyl group was then
converted into a carboxylic acid through two iterative oxidations
(IBX; NaClO2) to afford 48 in 63% overall yield from 47. This
newly unveiled motif was then enlisted to complete the
remaining structural elements of the target, namely, the B-, C-,
At this juncture, the completion of the target molecule (1)
required only a few finishing touches: the installation of the
(21) Wipf, P.; Miller, C. P. J. Org. Chem. 1993, 58, 3604-3606.
(22) A typical yield based on recovered starting material for this process was
around 50%.
(23) Because the majority of the substrate was reductively debrominated in the
Ph3SnH-induced reaction while the bulk of the remaining material from
the Witkop conditions remained unchanged from starting material, this
observation lends further support that this particular macrocyclization
reaction proceeds through a mechanism based on intramolecular photoin-
duced electron transfer from the D-ring indole chromophore to the adjacent
benzenoid E-ring.
(20) While we would be unable to accomplish this closure through cationic
chemistry, the anionic alternative does work (see ref 5).
9
12894 J. AM. CHEM. SOC. VOL. 126, NO. 40, 2004

Diazonamide A: First Total Synthesis
A R T I C L E S
Scheme 9. Elaboration of Advanced Intermediate 7 to 3, a Compound Bearing Both Macrocyclic Domains of the Revised Structure of
Diazonamide A (1)a
a Reagents and conditions: (a) (Boc)2O (30 equiv), aq NaHCO3/dioxane (1:2), 25 °C, 24 h, 91% based on recovered starting material; (b) IBX (3.0
equiv), DMSO, 25 °C, 2 h; (c) NaClO2 (5.0 equiv), NaH2PO4 (5.0 equiv), resorcinol (5.0 equiv), DMSO/H2O (10:1), 0f25 °C, 2 h, 66% over two steps;
(d) 6 (3.0 equiv), EDC (3.0 equiv), HOBt (3.0 equiv), NaHCO3 (15 equiv), DMF, 25 °C, 12 h, 65%; (e) POCl3/pyridine (1:4), 25 °C, 2 h, 52%; (f) hν (200
nm), epichlorohydrin (3.0 equiv), LiOAc (2.0 equiv), MeCN/H2O (3:1), 25 °C, 15 min, 33% or Ph3SnH (2.0 equiv), AIBN (0.1 equiv), C6H6, 80 °C, 3 h,
10%. IBX ) 1-hydroxy-1,2-benziodoxol-3(1H)-one, py ) pyridine, AIBN ) 2,2′-azobisisobutyronitrile.
Scheme 10. Final Stages and Completion of the First Total Synthesis of Diazonamide A (1)a
a Reagents and conditions: (a) NCS (4.0 equiv), CCl4/THF (1:1), 60 °C, 2 h, 53%; (b) TFA, 25 °C, 10 min, 98%; (c) DIBAL-H (1.0 M in toluene, 10
equiv, added portionwise), THF, -78f25 °C, 3 h, 56%; (d) H2 (2.0 atm), Pd(OH)2/C (20 wt %, catalytic), EtOH, 25 °C, 2 h; (e) 2 (5.0 equiv), EDC (5.0
equiv), HOBt (5.0 equiv), NaHCO3 (15 equiv), DMF, 25 °C, 12 h, 82% over two steps. NCS ) N-chlorosuccinimide.
two aryl chlorines, the formation of diazonamide A’s FH aminal
system, and the attachment of the final terminal residue (2-
hydroxyisovaleric acid). As shown in Scheme 10, the first of
these tasks was accomplished by treating 3 with NCS at 60 °C
for 2 h in a 1:1 mixture of THF and CCl4, conditions that led
to the site- and atropselective incorporation of the two requi-
site chlorines. Although our earlier chlorination forays1a using
just the BCD indole-oxazole subunit suggested that indole
protection was required for this event, we did not observe any
of the N-chlorination that plagued those studies in our reactions
with 3.
Seeking now to form the FH aminal system along the lines
employed previously (cf. Scheme 3), we deprotected the Boc
group from the G-ring phenol in 49 (TFA, 25 °C, 10 min) to
obtain 50, and then we exposed this new compound to a variety
of hydride sources. Amazingly, this cyclization proved quite
9
J. AM. CHEM. SOC. VOL. 126, NO. 40, 2004 12895

A R T I C L E S
Nicolaou et al.
challenging despite the ease by which a plethora of reducing
agents had managed to convert 24 into 27 earlier (see Table 2).
Following extensive scouting, we discovered that this task could
be accomplished in 56% yield through the portionwise addition
of 100 equiv of DIBAL-H (5 × 20 equiv) to a solution of 50
in THF over 3 h, using a cooling and warming cycle between
-78 °C and 25 °C after each addition.24 Although a seemingly
esoteric protocol, adding the hydride source in batches was
critical for the success of this cascade sequence, as otherwise
the reaction stalled and far reduced yields of 52 were consis-
tently obtained with starting material recovered instead. No other
reducing agent screened provided any product (52).
opportunities for discovery, whether as a source of inspiration
leading to creative strategies or as a testing ground revealing
weaknesses in the power of available methodology to effectively
fashion such complexity worthy of repair. Diazonamide A (1)
certainly is a natural product endowed with several such
domains, as the completion of the total synthesis described in
this article could not have been accomplished without the
development of several unique synthetic strategies and complex-
ity-building reaction cascades. Most noteworthy among these
are (1) the expeditious construction of the C-11 quaternary center
and its adjoining aromatic systems, (2) the successful use of
the Witkop-photocyclization reaction to forge the final bond
needed to close the highly hindered 12-membered heterocyclic
core, and (3) DIBAL-H initiated aminal formation.
Importantly, though, the story of the developed chemistry did
not stop once we reached this pinnacle. As the following article
will describe,28 not only did we use this sequence to prepare a
series of simpler analogues for biological evaluation, but we
also devised and executed an entirely different approach to the
target molecule that proved equally capable of handling its most
nefarious structural motifs.
At this stage, the spectral data for synthetic 52 and that
published25 for the natural product were quite similar, suggesting
that 1 was perhaps the correct structure for diazonamide A.
However, the moment of truth would not occur until the final
2-hydroxyisovaleric acid residue was attached to the molecule.
Thus, we chemoselectively excised the Cbz group guarding the
C-2 amine through a standard hydrogenation protocol using
Pearlman’s catalyst [20% Pd(OH)2/C] in EtOH and then
separately installed both (S)-2 and (R)-2 onto the resultant
product through a peptide coupling reaction as orchestrated by
EDC and HOBt. Of the two resultant materials, both formed in
82% overall yield from 52, one [the product obtained from the
use of (S)-2 in the coupling reaction] matched the data reported
Acknowledgment. We thank Drs. D. H. Huang and G.
Siuzdak for NMR spectroscopic and mass spectrometric as-
sistance, respectively. Financial support for this work was
provided by The Skaggs Institute for Chemical Biology, the
National Institutes of Health, predoctoral fellowships from the
National Science Foundation, Pfizer, and Bristol-Myers Squibb
(all to S.A.S.), a postdoctoral fellowship from the Skaggs
Institute for Chemical Biology (X.H.), and a grant from
American BioScience.
1
for natural diazonamide A exactly (on the basis of H NMR,
optical rotation, HPLC, and TLC comparisons in a variety of
solvent systems).26 As such, the first total synthesis of 1 was
complete through a route that required 21 steps in its longest
linear sequence, and all questions regarding the structural
identity of diazonamide A were finally laid to rest.27
Supporting Information Available: Experimental procedures
and compound characterization. This material is available free
of charge via the Internet at http://pubs.acs.org.
Conclusion
Unique and challenging molecular motifs within secondary
metabolites have long presented synthetic chemists with golden
JA040092I
(24) Neither of the conditions developed for our key reactions in our model
studies proved directly applicable to fully functionalized intermediates.
Although such outcomes often occur, its consistency throughout the
diazonamide campaign is indicative of the severely compact nature of the
target in which the alteration of a single group or the addition of a few
atoms can instigate several new problems with established chemistry.
(25) Lindquist, N.; Fenical, W.; Van Duyne, G. D.; Clardy, J. J. Am. Chem.
Soc. 1991, 113, 2303-2304.
(26) While most of the signals for both 1 and 53 were identical, a few were
diagnostic. Especially revealing was the C-37 proton signal (400 MHz,
MeOH-d4) which, in 1, coincided precisely with the reported value for the
natural product (3.88 ppm), whereas in 53 this proton appeared at 3.92
ppm.
(27) In line with expectations, synthetic diazonamide A (1) exhibited potent
cytotoxic activity at single digit nM concentrations against a variety of
human cancer cell lines of distinct origin, including ovarian carcinoma 1A9,
lung carcinoma A549, prostate carcinoma PC-3, breast carcinoma MCS-7,
and the Taxol-resistant 1A9/PTX10 cell line. We thank Dr. Paraskevi
Giannakakou and Aurora O’Brate, Emory University Winship Cancer
Center, for performing these studies. More detailed discussion and additional
biological studies are described in the following article.28
(28) Nicolaou, K. C.; Hao, J.; Reddy, M. V.; Bheema Rao, P.; Rassias, G.;
Snyder, S. A.; Huang, X.; Chen, D. Y.-K.; Brenzovich, W. E.; Giuseppone,
N.; Giannakakou, P.; O’Brate, A. J. Am. Chem. Soc. 2004, 126, XXXX-
XXXX.
9
12896 J. AM. CHEM. SOC. VOL. 126, NO. 40, 2004