Studies toward diazonamide A: Development of a hetero-pinacol macrocyclization cascade for the construction of the bis-macrocyclic framework of the originally proposed structure

Article abstract of DOI:10.1021/ja040091q

In this article, we describe further studies toward the originally proposed structure of diazonamide A (1). After confronting a number of failures in synthesizing the heterocyclic core of that structure, success was finally realized through the developmen

Full text of DOI:10.1021/ja040091q

Published on Web 07/27/2004
Studies toward Diazonamide A: Development of a
Hetero-Pinacol Macrocyclization Cascade for the Construction
of the Bis-Macrocyclic Framework of the Originally Proposed
Structure
K. C. Nicolaou,* Scott A. Snyder, Nicolas Giuseppone, Xianhai Huang, Marco Bella,
Mali V. Reddy, Paraselli Bheema Rao, Alexandros E. Koumbis,
Paraskevi Giannakakou,† and Aurora O’Brate†
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: In this article, we describe further studies toward the originally proposed structure of diazonamide
A (1). After confronting a number of failures in synthesizing the heterocyclic core of that structure, success
was finally realized through the development of a novel hetero-pinacol-based macrocyclization cascade
sequence. Subsequent elaboration led to an advanced compound bearing both of the 12-membered rings
of the target molecule. In addition, preliminary biological studies with intermediates and simplified analogues
obtained via the developed sequences are also described.
Introduction
to a point where some of the established chemistry can still be
employed. In assessing which of these options to pursue, we
In the preceding article in this issue,1 we delineated our first-
generation approach to the originally proposed structure of
diazonamide A (1, Scheme 1), one which proved capable of
delivering the entire heteroaromatic core of the molecule in
model systems but untenable in its late stages when applied to
fully functionalized intermediates. Fortunately, such challenges
often serve to extend the boundaries of established chemistry
by forcing the practitioner to develop more creative solutions
to unprecedented problems. As this article will detail, one such
answer did eventually become apparent in the form of a designed
and efficient cascade sequence using a hetero-pinacol-based
macrocyclization to forge the final C-C bond needed to close
the heterocyclic domain of 1 (see Scheme 1). This new approach
not only accelerated our studies to access the remainder of the
originally proposed structure of diazonamide A, work which
culminated in the construction of both macrocyclic units of 1,
but it also led to the invention of several new synthetic methods.
Equally important, it fueled initial chemical biology studies that
revealed a number of unexpected structure-activity relationships
within the diazonamide class.
felt that the first was far too extreme and instead anticipated
that the viability of our original strategy to diazonamide A (1)
could be resurrected by altering the manner in which we forged
the C29-C30 bond to close the heteroaromatic core. Indeed,
since the problem we had encountered with our first-gen-
eration approach was not the ring closure itself but obtaining
subsequent functionalization, maybe that issue could be cir-
cumvented if the needed functional groups were installed
directly during the cyclization step. For example, if an aldol-
type ring closure of compounds of type 3 could be accomplished,
with X representing a masked form of an amine such as a
phthalimide, cyanide, nitro group, or azide, then hopefully it
would be straightforward to form 2. From there, the A-ring
oxazole could be built either through an oxidation followed by
Robinson-Gabriel cyclodehydration, or oxazoline formation
followed by aromatization.
Apart from its appeal as an idea that would be easy to test
(since we would only need to construct different versions of
4), our failure to accomplish macrocyclic ring closures at other
sites during our model studies enhanced our desire to devote
time to its pursuit. For example, our attempts to form the 12-
membered heteroaromatic ring through the C16-C18 biaryl axis
both by intramolecular Ullmann-type2 and radical-based reac-
tions of compound 9 (see Scheme 2) met with resistance, even
though the test intermediate possessed an open F-ring to allow
for maximal flexibility.
Results and Discussion
1. Studies toward Alternate Macrocyclization. Anytime that
a strategy toward a complex molecule requires revision, one
faces the difficult decision of either scrapping the original
approach entirely, or retreating within the developed sequence
† Emory University School of Medicine, Winship Cancer Institute.
(1) Nicolaou, K. C.; Snyder, S. A.; Huang, X.; Simonsen, K. B.; Koumbis, A.
E.; Bigot, A. J. Am. Chem. Soc. 2004, 126, 10162-10173.
(2) Both classical, as well as milder, variants of this reaction were attempted.
For leading references, see: Hennings, D. D.; Iwama, T.; Rawal, V. H.
Org. Lett. 1999, 1, 1205-1208 and references therein.
9
10174
J. AM. CHEM. SOC. 2004, 126, 10174-10182
10.1021/ja040091q CCC: $27.50 © 2004 American Chemical Society

Studies toward Diazonamide A
A R T I C L E S
Scheme 2. Alternate Approaches To Close the Heterocyclic
Scheme 1. Revised Retrosynthetic Analysis of the Originally
Proposed Structure of Diazonamide A (1) Based on Alternate
Methods of Ring Closure
Macrocycle of Diazonamide A (1) Based on Different Sites of Ring
Closure Faila
a Reagents and conditions: (a) LiBH4 (2.0 M in THF, 2.0 equiv), Et2O,
0 °C, 2 h, 90%; (b) MeI (1.5 equiv), K2CO3 (3.0 equiv), DMF, 25 °C, 2 h,
94%; (c) Dess-Martin periodinane (2.0 equiv), NaHCO3 (10 equiv),
CH2Cl2, 0 °C, 2 h, 87%; (d) 8 (1.0 equiv), KHMDS (0.5 M in toluene, 1.0
equiv), THF, -78 °C, 5 min; 7 (1.0 equiv), -78 f 0 °C, 3 h, 54% (18%
9 + 36% trans isomer). TBS ) tert-butyldimethylsilyl; DMF ) N,N-
dimethylformamide; KHMDS ) potassium bis(trimethylsilyl)amide.
Scheme 3. Conversion of Indole 11 into Alternate Building Blocks
12, 13, and 14a
Accordingly, our renewed efforts to accomplish C29-C30
macrocyclization via the plan delineated in Scheme 1 began by
converting the alcohol within indole-oxazole 11 (Scheme 3) to
a series of masked amines. Thus, treatment of 11 with
phthalimide under Mitsunobu conditions3 led to the efficient
formation of 12 in 87% yield after just 3 h of reaction time at
0 °C. Similar smoothness was observed in the formation of azide
13 upon reaction of 11 with the pNO2 derivative4 of diphen-
ylphosphoryl azide. Finally, cyanide 14 was obtained over two
steps in 57% yield overall via initial halogen exchange followed
by nucleophilic displacement with cyanide from KCN as
promoted by 18-crown-6.
These three building blocks were then serially coupled to the
previously synthesized EFG boronate fragment 51 (see Scheme
4) using Suzuki coupling conditions5 described earlier [Pd(dppf)-
Cl2‚CH2Cl2, K2CO3, DME, 85 °C, 12 h]. Subsequent elaboration
of these intermediates (i.e., 15, 16, and 17) to the desired test
substrates (i.e., 18, 19, and 20) was then accomplished via HF-
mediated silyl ether cleavage, reformation of the acetonide that
broke apart in the previous reaction, and then oxidation (using
a Reagents and conditions: (a) phthalimide (1.2 equiv), Ph3P (1.2 equiv),
DEAD (1.1 equiv), THF, 0 °C, 3 h, 87%; (b) (pNO2Ph)2P(O)N3 (1.2 equiv),
DBU (1.2 equiv), toluene, 25 °C, 2 h, 83%; (c) Ph3P (1.7 equiv), imidazole
(2.0 equiv), I2 (1.5 equiv), CH2Cl2, 0 °C, 20 min, 94%; (d) KCN (2.0 equiv),
18-crown-6 (0.5 equiv), MeCN, 25 °C, 12 h, 61%. DEAD ) diethyl
azodicarboxylate; DBU ) 1,8-diazabicyclo[5.4.0]undec-7-ene; imid )
imidazole.
Swern conditions to access 18 and 19 and Dess-Martin periodi-
nane to generate 20). Unfortunately, despite the ease by which
these three substrates were prepared with the established chem-
istry, effecting their macrocyclization proved impossible. Ex-
posure of each to a variety of bases (NaH, KHMDS, LiHMDS,
(3) Gilbert, A. M.; Antane, M. M.; Argentieri, T. M.; Butera, J. A.; Francisco,
G. D.; Freeden, C.; Gundersen, E. G.; Graceffa, R. F.; Herbst, D.; Hirth,
B. H.; Lennox, J. R.; McFarlane, G.; Norton, N. W.; Quagliato, D.; Sheldon,
J. H.; Warga, D.; Wojdan, A.; Woods, M. J. Med. Chem. 2000, 43, 1203-
1214.
(4) Mizuno, M.; Shiori, T. Chem. Commun. 1997, 2165-2166.
(5) Ishiyama, T.; Murata, M.; Miyaura, N. J. Org. Chem. 1995, 60, 7508-
7510.
9
J. AM. CHEM. SOC. VOL. 126, NO. 32, 2004 10175

A R T I C L E S
Nicolaou et al.
Scheme 4. Attempts to Accomplish Anion-Based
phosphonate irreversibly drove our previously successful Hor-
ner-Wadsworth-Emmons (HWE) macrocyclizations,1,6 we
then turned to the synthesis of compound 24. Our expectation
was that Dieckmann condensation would overcome this prob-
lem,7 especially since the Vedejs group had reported8 success
in a related approach in their model studies toward 1. Unfor-
tunately, this advanced intermediate similarly proved recalcitrant
to macrocyclization under any conditions probed (including
NaH, LiHDMS, KOt-Bu, and NaOMe).
Macrocyclizationsa
2. Revised Retrosynthetic Analysis and Execution of the
New Strategy. Although this litany of failures was certainly
frustrating, it pushed us to think even more deeply about solving
the problem of C29-C30 functionalization. Mindful of our
previous use of the McMurry reaction (a pinacol coupling) to
form a highly strained eight-membered ring in our total synthesis
of Taxol,9 we hypothesized that perhaps we could enlist its
hetero variant to fashion a fully functionalized C29-C30 bond
for diazonamide A from a precursor aldehyde-oxime. This idea
seemed encouraging since intramolecular hetero-pinacol cou-
plings10 have been used on numerous occasions to fashion a
diverse range of rings ever since the late 1970s when the Corey,
Hart, and Bartlett groups independently demonstrated11 that
oximes could serve as competent radical acceptors. For example,
as shown in Part A of Scheme 5, the Naito group recently em-
12
ployed a hetero-pinacol cyclization initiated by SmI2 to effi-
ciently convert 25 into a seven-membered ring (26) appropriately
functionalized to complete a total synthesis of balanol (27).13
Despite this wealth of precedent, however, up to the end of
the year 2000, no variant of the hetero-pinacol reaction had been
successfully applied in a macrocyclization event to generate a
ring size greater than seven, despite precedent for medium-size
ring formation in related systems employing dialdehydes. In
fact, a number of studies seeking to form such rings met only
with failure.10a Nevertheless, we thought that the diazonamide
(6) Nicolaou, K. C.; Snyder, S. A.; Simonsen, K. B.; Koumbis, A. E. Angew.
Chem., Int. Ed. 2000, 39, 3473-3478.
(7) Numerous applications of this strategy in the synthesis of complex molecules
have appeared. For a representative survey, see: Davis, B. R.; Garratt, P.
J. In ComprehensiVe Organic Synthesis; Trost, B. M., Fleming, I., Eds.;
Pergamon: Oxford, 1991; Vol. 2, pp 806-829.
(8) Vedejs, E.; Zajac, M. A. Org. Lett. 2001, 3, 2451-2454. For the successful
application of the same approach to the revised structure of diazonamide
A, see: Vedejs, E.; Zajac, M. A. Org. Lett. 2004, 6, 237-240.
(9) Nicolaou, K. C.; Yang, Z.; Liu, J. J.; Ueno, H.; Nantermet, P. G.; Guy, R.
K.; Claiborne, C. F.; Renaud, J.; Couladouros, E. A.; Paulvannan, K.;
Sorensen, E. J. Nature 1994, 367, 630-633.
(10) (a) Ritte, D.; Hazell, R.; Skrydstrup, T. J. Org. Chem. 2000, 65, 5382-
5390. (b) Keck, G. E.; Wager, T. T.; Rodriquez, J. F. D. J. Am. Chem.
Soc. 1999, 121, 5176-5190. (c) Keck, G. E.; McHardy, S. F.; Murry, J.
A. J. Org. Chem. 1999, 64, 4465-4476. (d) Tormo, J.; Hays, D. S.; Fu, G.
C. J. Org. Chem. 1998, 63, 201-202. (e) Keck, G. E.; Wager, T. T. J.
Org. Chem. 1996, 61, 8366-8367. (f) Kiguchi, T.; Tajiri, K.; Ninomiya,
I.; Naito, T.; Hiramatsu, H. Tetrahedron Lett. 1995, 36, 253-256. (g)
Camps, P.; Font-Bardia, M.; Munoz-Torrero, D.; Solans, X. Liebigs Ann.
1995, 523-535. (h) Shono, T.; Kise, N.; Fujimoto, T.; Yamanami, A.;
Nomura, R. J. Org. Chem. 1994, 59, 1730-1740. (i) Naito, T.; Tajiri, K.;
Harimoto, T.; Ninomiya, I.; Kiguchi, T. Tetrahedron Lett. 1994, 35, 2205-
2206. (j) Marco-Contelles, J.; Martinez, L.; Martinez-Grau, A.; Pozuelo,
C.; Jimeno, M. L. Tetrahedron Lett. 1991, 32, 6437-6440.
a Reagents and conditions: (a) 5 (1.1 equiv), 12, 13, or 14 (1.0 equiv),
Pd(dppf)Cl2 (0.2 equiv), K2CO3 (5.0 equiv), DME, 85 °C, 12 h; (b) aq HF
(48%, excess), MeCN, 0 °C, 45 min; (c) 2,2-DMP, acetone, 25 °C, 5 min;
(d) DMSO (10 equiv), (COCl)2 (5.0 equiv), CH2Cl2, -78 °C, 45 min; Et3N
(20 equiv), CH2Cl2, -78 °C, 15 min or Dess-Martin periodinane (3.0
equiv), NaHCO3 (10 equiv), CH2Cl2, 25 °C, 1 h, 51% overall for 18, 10%
overall for 19, 40% overall for 20. dppf ) diphenylphosphinoferrocene;
DME ) ethylene glycol dimethyl ether; 2,2-DMP ) 2,2-dimethoxypropane.
(11) (a) Corey, E. J.; Pyne, S. G. Tetrahedron Lett. 1983, 24, 2821-2824. (b)
Hart, D. J.; Seely, F. L. J. Am. Chem. Soc. 1988, 110, 1631-1633. (c)
Bartlett, P. A.; McLaren, K. L.; Ting, P. C. J. Am. Chem. Soc. 1988, 110,
1633-1634.
(12) (a) Namy, J. L.; Girard, P.; Kagan, H. B. NouV. J. Chim. 1977, 1, 5-7. (b)
Girard, P.; Namy, J. L.; Kagan, H. B. J. Am. Chem. Soc. 1980, 102, 2693-
2698. For recent review articles on the application of SmI2 in organic
synthesis, see: (c) Kagan, H. B. Tetrahedron 2003, 59, 10351-10372. (d)
Krief, A.; Laval, A.-M. Chem. ReV. 1999, 99, 745-778. (e) Molander, G.
A.; Harris, C. R. Tetrahedron 1998, 54, 3321-3354. (f) Skrydstrup, T.
Angew. Chem., Int. Ed. Engl. 1997, 36, 345-347. (g) Molander, G. A.;
Harris, C. R. Chem. ReV. 1996, 96, 307-338.
LDA, or KOt-Bu) universally failed to deliver anything
resembling the desired product. Instead, we typically observed
decomposition and, in some cases, recovered starting material.
Recognizing that the failures in these couplings could reflect
the high proclivity for retro-aldol reactions, since nothing existed
to prevent such an outcome in the same way that the loss of a
(13) Miyabe, H.; Torieda, M.; Inoue, K.; Tajiri, K.; Kiguchi, T.; Naito, T. J.
Org. Chem. 1998, 63, 4397-4407.
9
10176 J. AM. CHEM. SOC. VOL. 126, NO. 32, 2004

Studies toward Diazonamide A
A R T I C L E S
Scheme 5. Selected Uses of SmI2 in Organic Synthesis as Pertinent Precedents to the Diazonamide Problem
Scheme 6. Revised Retrosynthetic Analysis of the Originally Proposed Structure of Diazonamide A (1) Based on a Heteropinacol Coupling
Sequence
framework might provide a particularly unique case of a 12-
membered ring, as π-π stacking between the B- and E-rings
could bring the aldehyde and oxime motifs quite close and
significantly reduce their rotational freedom. Thus, the reaction
might have at least some chance for success, although conven-
tional wisdom seemed to point to a challenging proposition.
productively combine this transformation into a new reaction
cascade that could potentially have wider applications.
Our explorations of this strategy began with the synthesis of
a new EFG building block, as shown in Scheme 7, altered only
from the fragment employed earlier (i.e., 5) by the use of a
methyl ester as the methylene activating group in the 5-exo-tet
cyclization leading to 36. Although the yield for this key
conversion was lower (38%) than that achieved previously1 with
cyanide due to a number of side-reactions initiated by the ester
functionality, the presence of that motif provided a reactive
handle easily convertible to a C-11 lactol and incapable of
benzofuran fragmentation. As an added benefit, its incorporation
also shortened the overall sequence from L-tyrosine methyl ester
to the final boronate ester (37) by three steps.
Since we would ultimately need an amide product, and not
an N-O bond as that observed in the formation of compound
26 (see Scheme 5), following cyclization, we would have to
break that linkage apart. As shown in Part B of Scheme 5, we
were pleased to discover recent reports from the Keck group
detailing that SmI2 is particularly effective in this task.14 Thus,
it appeared to us that it should be possible to accomplish both
hetero-pinacol coupling and N-O cleavage in the same pot.
Although a seemingly safe assumption, an extensive search of
the literature revealed its success on only one occasion through
a protocol which involved treating a substrate with 6.0 equiv
of SmI2 and then stirring for a prolonged period in deoxygenated
H2O.15,16 Our analysis of this unique result suggested that its
success reflected the amount of SmI2 employed, since if 4-5
equiv of SmI2 were required to accomplish just one of the
reactions in Scheme 5, then at least 8-10 equiv would probably
be required to effect both. Accordingly, our new plan for the
synthesis of diazonamide A was to treat an intermediate such
as 33 (see Scheme 6) with a gross excess of SmI2, hoping that
both cyclization and N-O cleavage could be accomplished
concomitantly to afford a 1,2-amino alcohol product. Rather
than isolate that likely polar intermediate, we would then try
and couple that product directly with a protected form of L-valine
to obtain 32. Thus, if this sequence could be realized, not only
would it constitute the first example of a hetero-pinacol reaction
leading to a medium- or large-sized ring, but it also would
Once this fragment was completed, it was coupled with the
previously obtained indole-oxazole 381 through the standard
Miyaura conditions5 for Suzuki coupling of boronate esters,
leading to the assembly of 39 (see Scheme 8) in 83% yield.
This new intermediate was then cleanly converted into aldehyde-
oxime 33 through a tandem deprotection/oxidation sequence,
followed by selective oxime capture of the more-activated
(14) (a) Keck, G. E.; Wager, T. T.; McHardy, S. F. Tetrahedron 1999, 55,
11755-11772. (b) Chiara, J. L.; Destabel, C.; Gallego, P.; Marco-Contelles,
J. J. Org. Chem. 1996, 61, 359-360. (c) Keck, G. E.; McHardy, S. F.;
Wager, T. T. Tetrahedron Lett. 1995, 36, 7419-7422.
(15) (a) Marco-Contelles, J.; Gallego, P.; Rodr´ıguez-Ferna´ndez, M.; Khiar, N.;
Destabel, C.; Bernabe´, M.; Mart´ınez-Grau, A.; Chiara, J. L. J. Org. Chem.
1997, 62, 7397-7412. For a preliminary disclosure of the same reductive
coupling/N-O cleavage and an example with altered protecting groups,
see (b) Chiara, J. L.; Marco-Contelles, J.; Khiar, N.; Gallego, P.; Destabel,
C.; Bernabe´, M. J. Org. Chem. 1995, 60, 6010-6011. (c) Bobo, S.; Storch
de Gracia, I.; Chiara, J. L. Synlett 1999, 1551-1554.
(16) The role of H2O is either as a proton source, or, more likely, as a donor
ligand which increases the reducing power of SmI2. For leading references,
see: (a) Hanessian, S.; Girard, C. Synlett 1994, 861-862. (b) Hasegawa,
E.; Curran, D. P. J. Org. Chem. 1993, 58, 5008-5010.
9
J. AM. CHEM. SOC. VOL. 126, NO. 32, 2004 10177

A R T I C L E S
Nicolaou et al.
Scheme 7. Synthesis of Advanced EFG Fragment 37a
an alternative to either of these pictures, one could also invoke
a samarium-bridged diradical such as 45 to account for the
eventual formation of the bridged bond in 41.
Apart from these mechanistic considerations, it is worth
noting that, in accordance with earlier reports exploring the
hetero-pinacol coupling of aldehyde-oximes,10 the macrocy-
clization of 33 did not proceed in the absence of HMPA.
Moreover, if the ratio of HMPA/SmI2 was reduced from 4:1 to
2:1 (still using 9 equiv of SmI2), 32 was observed, along with
significant amounts of cyclized product, with the N-O linkage
firmly intact, indicating that the presence of a suitable donor
ligand in conjunction with excess SmI2 is the combination
required for reliable oxime cleavage following reductive cy-
clization. These results suggest that in cases where one would
desire to effect only N-O cleavage, the addition of HMPA
might greatly facilitate the transformation in circumstances that
prove difficult or low-yielding with SmI2 alone.
Having finally established a functionalized C29-C30 bridge
after numerous failures, we could now investigate means by
which to form the A-ring oxazole and complete the entire het-
eroaromatic core of 1. As mentioned earlier, two direct pathways
were available to accomplish this goal from advanced interme-
diate 32: oxidation followed by Robinson-Gabriel dehydration,
or oxazoline formation and subsequent aromatization. While
both might appear equally feasible on paper, our model studies
suggested that the first was unlikely to succeed with 32 since
only pTsOH in refluxing toluene proved effective in initiating
the needed cyclodehydration on a substrate that lacked its
numerous acid-labile functionalities.18 Accordingly, we elected
to probe oxazoline formation and found that the desired ring
system (i.e., 46, Scheme 9) could indeed be formed in 24%
yield over the course of 12 h using (diethylamino)sulfur
trifluoride (DAST) in THF at -10 °C.19 The remaining material
balance from this event was the enamide that resulted from
simple dehydration of the C-30 alcohol. Although the yield for
this operation was low, that outcome partially reflects the fact
that only one of the two possible oxazoline products (which
we have tentatively assigned as drawn in Scheme 9) was formed
in the event.20 Consequently, only the portion of starting material
that possessed the proper C-29 and C-30 stereochemistry
required to form this product could have participated produc-
tively in the cyclization. In truth, the efficiency of this reaction
(or lack thereof) was ultimately of no consequence since we
could never accomplish the subsequent oxidation leading to 31,
despite nearly a dozen attempts.21
a Reagents and conditions: (a) BrCH2CO2Me (1.5 equiv), K2CO3 (1.5
equiv), acetone, 56 °C, 3 h, 95%; (b) 2,2-dimethoxypropane (10 equiv),
pTsOH (0.1 equiv), acetone, 25 °C, 30 min, 87%; (c) mCPBA (77%, 2.0
equiv), NaHCO3 (3.0 equiv), CH2Cl2, 0 °C, 8 h, 84%; (d) tBuOK (1.0 M
in THF, 1.5 equiv), THF/DMF (1:1), -78 °C, 5 min, then 1 M aq HCl,
-78 °C, 5 min, 38%; (e) TBSCl (3.0 equiv), imidazole (6.0 equiv), DMF,
25 °C, 6 h, 88%; (f) bis(pinacolato)diboron (1.1 equiv), Pd(dppf)Cl2 (0.2
equiv), KOAc (3.0 equiv), DMSO, 85 °C, 6 h, 81%. pTsOH ) p-
toluenesulfonic acid; mCPBA ) m-chloroperoxybenzoic acid; BPD )
bis(pinacolato)diboron.
aldehyde after just 10 min of reaction with MeONH2‚HCl in
DMSO at 25 °C. Use of neat DMSO as solvent was critical for
the success of this final operation leading to 33, as the same
reaction in any alcoholic media was attended by a significant
degree of acetonide cleavage (likely due to trace HCl derived
from the oxime source). In any case, with an efficient synthesis
of 33 achieved in 70% yield from 39, explorations into the key
reaction of the proposed sequence could begin. Most gratify-
ingly, little reaction scouting was required. Following treatment
of aldehyde-oxime 33 with a premixed complex of 9 equiv of
freshly prepared SmI2 and 36 equiv of HMPA in THF at 25 °C
for 1 h, followed by an aqueous NH4Cl reaction quench,
extraction, solvent removal, and subsequent peptide coupling
using a DMF solution of Fmoc-protected L-valine (44), EDC,
and HOBt, we obtained compound 32 as a mixture of stereo-
isomers in an isolated yield of 45%.
Mechanistically, we believe that the initial exposure of 33 to
SmI2/HMPA led to the generation of diradical intermediate 40,
which then cyclized to provide 41. The presence of excess SmI2
complexed with HMPA then effected N-O cleavage, leading
to intermediate 42, which, upon workup, provided the desired
amino alcohol that was later trapped through peptide formation.
Consequently, each step in the sequence proceeded in an average
yield of 75%. Although this picture differs from the more typical
representation17 of this reaction where a ketyl radical attacks
an intact oxime to account for the generation of 41, the isolation
of noncyclized material with both the aldehyde and oxime
reduced suggests that the existence of diradical 40 cannot be
excluded. We also believe that our alternative takes into account
the unique framework of diazonamide A (1), as our experience
with compounds such as 33 suggests that the steric hindrance
around that aldehyde would make it quite difficult to form a
ketyl radical without touching the highly accessible oxime. As
With this approach reaching a roadblock, we turned to the
second alternative for oxazole formation, and in line with our
previous expectations, Robinson-Gabriel cyclodehydration
proved impossible to achieve following the formation of 47
either with Dess-Martin periodinane on small scale or TPAP/
(17) For example, see: Robertson, G. M. In ComprehensiVe Organic Synthesis;
Trost, B. M., Fleming, I., Eds.; Pergamon: Oxford, 1991; Vol. 3, pp 563-
611.
(18) Similar findings were presented in another diazonamide study: Wipf, P.;
Methot, J.-L. Org. Lett. 2001, 3, 1261-1264.
(19) Lafargue, P.; Guenot, P.; Lellouche, J.-P. Heterocycles 1995, 41, 947-
958.
(20) This assignment is based, in part, on the fact that the alternate oxazoline
would suffer from severe steric interactions with the protons from the C-11
position in ring F. Such strain also renders the approach of DAST or any
other reagent from that side of the molecule quite unlikely.
(21) Efforts to accomplish this conversion included several known oxidants for
the process, including: MnO2 in refluxing benzene or CH2Cl2, DDQ in
benzene at 25 °C, Pd/C in benzene at 25 °C, BrCCl3 and DBU in MeCN
at 25 °C, and NCS in CCl4 at 25 °C and reflux.
9
10178 J. AM. CHEM. SOC. VOL. 126, NO. 32, 2004

Studies toward Diazonamide A
A R T I C L E S
Scheme 8. Successful Generation of the Fully Functionalized Heterocyclic Core (33) of the Originally Proposed Structure of Diazonamide A
(1) Using Suzuki and Hetero-Pinacol Couplingsa
a Reagents and conditions: (a) 37 (1.1 equiv), 38 (1.0 equiv), Pd(dppf)Cl2 (0.2 equiv), K2CO3 (5.0 equiv), DME, 85 °C, 12 h, 83%; (b) TBAF (4.0 equiv),
THF, 25 °C, 20 min, 93%; (c) Dess-Martin periodinane (3.0 equiv), NaHCO3 (10 equiv), CH2Cl2, 25 °C, 1 h, 87%; (d) MeONH2‚HCl (5.0 equiv), DMSO,
25 °C, 10 min, 87%; (e) SmI2 (0.1 M in THF, 9.0 equiv), HMPA (36 equiv), THF, 25 °C, 1 h; then saturated aq NaHCO3, 25 °C, 1 h; solvent removal; then
44 (3.0 equiv), EDC (3.0 equiv), HOBt (3.0 equiv), DMF, 25 °C, 10 h, 40-45% overall, ca. 75% per synthetic operation in the designed reaction sequence.
TBAF ) tetra-n-butylammonium fluoride; HMPA ) hexamethylphosphoramide; EDC ) 3-(3-dimethylaminopropyl)-1-ethylcarbodiimide; HOBt ) 1-hydroxy-
1H-benzotriazole.
Scheme 9. Formation of the A-ring Oxazole Subunit of the Proposed Structure of Diazonamide A (1) from Advanced Intermediate 32a
a Reagents and conditions: (a) DAST (5.0 equiv), THF, -10 °C, 12 h, 24%; (b) Dess-Martin periodinane (3.0 equiv), NaHCO3 (10 equiv), CH2Cl2, 25
°C, 1 h, 83% or TPAP (1.0 equiv), NMO (5.0 equiv), CH2Cl2, 25 °C, 2 h, 68%; (c) POCl3/pyridine (1:2), 70 °C, 6 h, 45%, 68% based on recovered starting
material. DAST ) (diethylamino)sulfur trifluoride; TPAP ) tetra-n-propylammonium perruthenate; NMO ) 4-methylmorpholine N-oxide; py ) pyridine.
NMO (TPAP ) tetra-n-propylammonium perruthenate, NMO
) 4-methylmorpholine N-oxide) on larger scale. As shown in
Table 1, no published oxazole formation protocol22 proved
successful in converting 47 into 31, with some providing
9
J. AM. CHEM. SOC. VOL. 126, NO. 32, 2004 10179

A R T I C L E S
Nicolaou et al.
Table 1. Screening of Conditions To Accomplish the Formation of
the A-ring Oxazole and Complete the Heterocyclic Core of
Diazonamide A
Scheme 10. Completion of the Second Macrocyclic Subunit of the
Originally Proposed Structure of Diazonamide A (1)a
entry
R
conditions
yield
1
2
3
4
5
6
7
8
Fmoc Burgess reagent, THF, ∆, 10 h
no reaction
Fmoc Ph3P, I2, Et3N, Ch2Cl2, 25 °C, 10 h no reaction
Fmoc Ph3P, Cl3CCCl3, Et3N, 25 °C, 10 h no reaction
Fmoc TFAa, TFAAb, 60 °C, 2 h
Fmoc POCl3, 70 °C, 2 h
Fmoc P2O5, CHCl3, 25 °C, 5 h
Fmoc POCl3/pyridine (1:2), 70 °C, 6 h
decomposition
decomposition
decomposition
45% 31 (33% rsm)
34% 49 (20% rsm)
Boc
POCl3/pyridine (1:2), 70 °C, 6 h
a TFA ) trifluoroacetic acid. b TFAA ) trifluoroacetic anahydride.
recovered starting material (entries 1-3) and others leading to
complete decomposition (entries 4-6). Use of the same
protocols with microwave activation23 instead of direct heat or
efforts to minimize the steric bulk of the valine residue through
alternate nitrogen protection (such as Boc or Alloc instead of
Fmoc, as well as the use of N3 for the entire amine) provided
no improvement either. Even probing the formation of a far
simpler oxazole by using an acetate in lieu of the L-valine
attached to the amine in 47 failed to afford any desired product.
Our approach was seemingly at another dead end.
After examining the collected results more carefully, we
wondered if a new opportunity for success might arise simply
by modifying an existing procedure to accomplish Robinson-
Gabriel cyclodehydration. More specifically, although POCl3
has proven to be quite effective at oxazole synthesis in a number
of contexts,22a perhaps when we attempted that reaction with
47, trace amounts of HCl from the reagent led its functionality
to break apart before oxazole formation could occur. Thus, if
we buffered that dehydrating reagent with a base such as
pyridine, then maybe this problem could be circumvented. As
an added benefit, it seemed reasonable to expect that rendering
the reaction media basic could also accelerate the overall rate
of the reaction by promoting the amide enolization required to
displace the POCl3-activated alcohol. When this protocol was
put to the test by adding a 2:1 mixture of pyridine/POCl3 to
ketone 47 and heating the resultant solution at 70 °C for 6 h,
we were pleased to discover that oxazole 31 could be obtained
in 45% yield with virtually all of the remaining material balance
(33%) constituting recovered starting material (Table 1, entry
7). Nearly equal levels of success were observed with related
substrates bearing alternate L-valine protection (entry 8).24
Having finally overcome this major synthetic hurdle, we felt
that the targeted structure (1) would soon be within reach
a Reagents and conditions: (a) aq HF (48%, excess), MeCN, 0 °C, 45
min, 96%; (b) Dess-Martin periodinane (3.0 equiv), NaHCO3 (10 equiv),
CH2Cl2, 25 °C, 1 h; (c) NaClO2 (5.0 equiv), NaH2PO4 (5.0 equiv), resorcinol
(5.0 equiv), DMSO/H2O (10:1), 25 °C, 3 h, 65% over two steps; (d) Et2NH
(excess), THF, 25 °C, 4 h, 97%; (e) HATU (2.0 equiv), collidine (6.0 equiv),
DMF/CH2Cl2 (9:1, 1.0 × 10-4 M), 25 °C, 12 h, 5-10%. HATU ) 2-(1H-
9-azabenzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate.
following formation of the second macrocyclic subunit through
lactamization, a transformation for which we anticipated few
difficulties on the basis of the wealth of literature precedent to
effect such ring closures. Indeed, our initial forays along these
lines were encouraging, as 31 was converted into 50 (see
Scheme 10) without incident in 61% overall yield through HF-
mediated cleavage of the acetonide, generation of a free acid
from the resultant alcohol through a two-stage oxidation
protocol, and finally, Et2NH-induced lysis25 of the Fmoc group
guarding the amine appended to the A-ring oxazole. Unfortu-
nately, as revealed in Table 2, initial explorations to effect the
formation of 51 from this advanced amino acid (50) met with
significant resistance (entries 1-4),26 leading in all cases either
to decomposition or dimerization (even when the reaction was
(23) Brain, C. T.; Paul, J. M. Synlett 1999, 1642-1644.
(24) The scope of this new protocol is explored more fully in the final article
in this series and has proven applicable to the formation of diverse oxazoles
from ketoamides, thiazolines from thioamide-alcohols, thiazoles from
ketothioamides, and furans from 1,4-dicarbonyls.
(22) Examples of protocols employed include: (a) Dow, R. L. J. Org. Chem.
1990, 55, 386-388. (b) Wasserman, H. H.; Vinick, F. J. J. Org. Chem.
1973, 38, 2407-2408. (c) Wipf, P.; Miller, C. P. J. Org. Chem. 1993, 58,
3604-3606. (d) Parsons, R. L.; Heathcock, C. H. J. Org. Chem. 1994, 59,
4733-4734. (e) Fukuyama, T.; Xu, L. J. Am. Chem. Soc. 1993, 115, 8449-
8450.
(25) The same operation could also be accomplished in slightly lower yield
using TBAF. For an application of this alternative protocol in total synthesis,
see: Jiang, W.; Wanner, J.; Lee, R. J.; Bounard, P.-Y.; Boger, D. L. J.
Am. Chem. Soc. 2002, 124, 5288-5290.
9
10180 J. AM. CHEM. SOC. VOL. 126, NO. 32, 2004

Studies toward Diazonamide A
A R T I C L E S
Table 2. Screening of Conditions To Accomplish Peptide
Formation and Thereby Complete Both Macrocycles of
Diazonamide A
entry
conditions
yield (%)
1
2
3
4
5
6
7
PyBroP, NaHCO3, DMF/CH2Cl2 (1:1)a
FDPP, NaHCO3, DMF/CH2Cl2 (1:1)a
EDC, HOBt, DMF/CH2Cl2 (1:1)a
dimer/trimer
decomposition
dimer/trimer
decomposition
∼60 of 52
trace
DEPBT, i-Pr2NEt, DMF/CH2Cl2 (1:1)a
DPPA, NaHCO3, DMF/CH2Cl2 (1:1)a
HATU, collidine, DMF/CH2Cl2 (1:1)a
HATU, collidine, DMF/CH2Cl2 (9:1)b
5-10c
Figure 1. Structural reassignment of diazonamide A (1) to 53 based on
the work of Harran and co-workers.30
a Concentration of 50 ) 5.0 × 10-4 M. b Concentration of 50 ) 1.0 ×
10-4 M. c Range indicates maximum and minimum values obtained for
several runs. PyBroP ) bromotripyrrolidinophosphonium hexafluorophos-
phate; FDPP ) pentafluorophenyl diphenylphosphinate; DEPBT ) 3-di-
ethyloxyphophoryloxy)-1,2,3-benzotriazin-4(3H)-one; DPPA ) diphen-
ylphosphoryl azide.
(51) via column chromatography. Despite the disappointing
outcome for this key reaction in terms of yield, the fact that it
could be achieved marked a milestone for our overall plan to
synthesize 1, especially since we felt that its outcome could be
improved through optimization. Sadly, though, we would never
get that opportunity since it was at this juncture that the original
structure of diazonamide A (1) was revised to 53 (see Figure
1) by the Harran group,30 leading us to abandon any further
synthetic efforts toward the “oxygen analogue” of the real
diazonamide A.
3. Analogue Synthesis and Chemical Biology Explorations.
While we would not complete the synthesis of the originally
proposed structure of diazonamide A (1), the developed
sequences did allow us to make a few preliminary forays into
the chemical biology of the diazonamide class. Such explorations
were certainly of value, even though a key part of our
intermediates did not exactly match the structure now proposed
for the natural product, since the Harran team had disclosed
that advanced synthetic materials bearing the central oxygen
atom of 1 could induce the same phenotype as natural di-
azonamide A in several tumor cell lines.30b Accordingly, we
not only screened every synthetic compound that we had in hand
as part of our program to synthesize 1, but we also constructed
and tested a series of simplified analogues. The preparation of
these substances is shown in Schemes 11 and 12.31
run at exceedingly low concentration). Although molecular
models suggested that the reactive units should be quite close,
the steric hindrance within the system and the entropic penalties
associated with the formation of the highly strained 12-
membered ring must have been the culprits that led to such
difficulties. Indeed, the only smooth reaction which we ever
observed in our first attempts occurred when 50 was exposed
to diphenylphosphoryl azide (DPPA)27 in NaHCO3-buffered
CH2Cl2, conditions which led to the formation of a product
which we have assigned as the more flexible 13-membered
cyclic urea 52 (the product of a room-temperature Curtius
rearrangement).28
Thankfully, after more experimentation, we were able to form
the desired macrolactam (51) in yields of 5-10% if 50 was
added dropwise over several hours to a mixture of 2-(1H-9-
azabenzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluoro-
phosphate (HATU) and 2,4,6-collidine dissolved in DMF/
CH2Cl2 (9:1) at a final concentration of 1.0 × 10-4 M.29 The
remainder of the material from this experiment was dimeric in
nature, and could not be separated from the desired product
As revealed in Figure 2, with selected data for a few of the
compounds examined, these studies unveiled a number of
interesting structure-activity relationships. For instance, al-
though analogues 60, 62, 63, and 64 are all several-hundred-
(29) For other 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.
(30) (a) Li, J.; Jeong, S.; Esser, L.; Harran, P. G. Angew. Chem., Int. Ed. 2001,
40, 4765-4769. (b) Li, J.; Burgett, A. W. G.; Esser, L.; Amezcua, C.;
Harran, P. G. Angew. Chem., Int. Ed. 2001, 40, 4770-4773.
(31) For a detailed discussion of this sequence, see: 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. For the sake of compari-
son, however, it is interesting to note here the decreased yield observed
for the key SmI2/HMPA-based hetero-pinacol sequence in this model system
versus the fully functionalized intermediates presented in Scheme 8.
(26) DEPBT: Li, H.; Jiang, X.; Ye, Y.-h.; Fan, C.; Romoff, T.; Goodman, M.
Org. Lett. 1999, 1, 91-93. FDPP: Chen, S.; Xu, J. Tetrahedron Lett. 1991,
32, 6711-6714. PyBrop: Coste, J.; Frerot, E.; Jouin, P. J. Org. Chem.
1994, 59, 2437-2446.
(27) Qian, L.; Sun, Z.; Deffo, T.; Mertes, K. B. Tetrahedron Lett. 1990, 31,
6469-6472.
(28) To achieve such a smooth rearrangement at a low temperature typically
requires the presence of a protic or Lewis acid; in this case, the particular
spatial arrangement of the reactive groups in 17 must facilitate this
conversion as no such activators were needed.
9
J. AM. CHEM. SOC. VOL. 126, NO. 32, 2004 10181

A R T I C L E S
Nicolaou et al.
Scheme 11. Execution of the Developed Pinacol Coupling
Cascade Sequence to Efficiently Prepare Ketoamide 59a
Scheme 12. Generation of a Series of Advanced Model System
Analogs Using the Developed Chemistrya
a Reagents and conditions: (a) POCl3, 70 °C, 2 h, 60/61 (2.4:1), 65%;
(b) aq NaOH (15%, excess), THF, 25 °C, 10 min, 74%; (c) NCS (3.0 equiv),
THF/CCl4 (1:1), 55 °C, 8 h, 73%; (d) BBr3 (1.0 M in CH2Cl2, 2.0 equiv),
CH2Cl2, -78 °C, 20 min; then aq NaOH (15%, excess), THF, 25 °C, 10
min, 61%. NCS ) N-chlorosuccinimide.
a Reagents and conditions: (a) 54 (1.0 equiv), 55 (1.0 equiv), Pd(dppf)Cl2
(0.2 equiv), K2CO3 (5.0 equiv), DME, 110 °C, 8 h, 66%; (b) TBAF (3.0
equiv), THF, 25 °C, 10 min, 93%; (c) Dess-Martin periodinane (3.0 equiv),
NaHCO3 (10.0 equiv), CH2Cl2, 25 °C, 1 h, 88%; (d) MeONH2‚HCl (10.0
equiv), DMSO, 25 °C, 10 min, 91%; (e) SmI2 (0.1 M in THF, 9.0 equiv),
HMPA (36 equiv), THF, 25 °C, 1 h; then saturated aq NH4Cl, 25 °C,1 h;
solvent removal; then AcOH (3.0 equiv), EDC (3.0 equiv), HOBt (3.0
equiv), DMF, 25 °C, 1 h, 25% overall, 71% per synthetic operation in the
cascade sequence; (f) Dess-Martin periodinane (3.0 equiv), NaHCO3 (10.0
equiv), CH2Cl2, 25 °C, 1.5 h, 94%.
fold less cytotoxic than diazonamide A (53) against the 1A9
human ovarian carcinoma cell line, these four simple structural
congeners indicate the importance of the two aryl chlorine
residues (60 vs 63 and 62 vs 64) in conferring biological efficacy
and that substitution of the indole nucleus (63 vs 64) can
potentially lead to increased potency. Apart from these pre-
liminary findings, however, we were unable to identify any
intermediate or analogue possessing potency remotely com-
mensurate to diazonamide A, even among compounds bearing
both of the macrocyclic subunits of 1.
Figure 2. Preliminary biological screening of simple heterocyclic core
analogues (60, 62, 63, and 64) against 1A9 human ovarian carcinoma cells.
in this series will detail,32 the developed chemistry proved
directly translatable into a successful total synthesis of the
revised structure of diazonamide A (53).
Acknowledgment. We thank Drs. D. H. Huang and G.
Siuzdak for NMR spectroscopic and mass spectrometric assis-
tance, respectively. Financial support for this work was provided
by The Skaggs Institute for Chemical Biology, the National
Institutes of Health (USA), predoctoral fellowships from the
National Science Foundation, Pfizer, and Bristol-Myers Squibb
(all to S.A.S.), postdoctoral fellowships from the Association
pour la Recherche sur le Cancer (N.G.) and The Skaggs Institute
for Chemical Biology (X.H.), and a grant from American
Biosciences.
Conclusion
Despite the interruption of the sequence developed to access
the originally proposed structure for diazonamide A (1), the
campaign waged to reach advanced intermediate 51 was far from
fruitless. Indeed, the adopted strategy inspired a novel extension
of the hetero-pinacol reaction as part of a unique cascade
sequence to construct complex macrocyclic systems. It also led
to the discovery of a series of valuable synthetic methodologies,
the most important in this article being the identification of a
powerful method to accomplish Robinson-Gabriel cyclodehy-
dration in hindered settings using POCl3/pyridine. In addition,
biological screening of analogues and intermediates synthesized
as part of this program revealed several significant findings,
particularly in terms of structure-activity relationships for the
diazonamide class. Most important, as the remaining two articles
Supporting Information Available: Experimental procedures
and compound characterization. This material is available free
of charge via the Internet at http://pubs.acs.org.
JA040091Q
(32) (a) Nicolaou, K. C.; Chen, D. Y.-K.; Huang, X.; Ling, T.; Bella, M.; Snyder,
S. A. J. Am. Chem. Soc. 2004, 126, in press. (b) 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, in press.
9
10182 J. AM. CHEM. SOC. VOL. 126, NO. 32, 2004