The first total synthesis of dragmacidin D

Article abstract of DOI:10.1021/ja027822b

The first total synthesis of the biologically significant bis-indole alkaloid dragmacidin D (5) has been achieved. Thermal and electronic modulation provides the key for a series of palladium-catalyzed Suzuki cross-coupling reactions that furnished the core structure of the complex guanidine- and aminoimidazole-containing dragmacidins. Following this crucial sequence, a succession of meticulously controlled final events was developed leading to the completion of the natural product.

Full text of DOI:10.1021/ja027822b

Published on Web 10/05/2002
The First Total Synthesis of Dragmacidin D
Neil K. Garg, Richmond Sarpong, and Brian M. Stoltz*
Contribution from The Arnold and Mabel Beckman Laboratories of Chemical Synthesis,
DiVision of Chemistry and Chemical Engineering, California Institute of Technology,
1200 East California BouleVard, MC 164-30, Pasadena, California 91125
Received July 23, 2002
Abstract: The first total synthesis of the biologically significant bis-indole alkaloid dragmacidin D (5) has
been achieved. Thermal and electronic modulation provides the key for a series of palladium-catalyzed
Suzuki cross-coupling reactions that furnished the core structure of the complex guanidine- and
aminoimidazole-containing dragmacidins. Following this crucial sequence, a succession of meticulously
controlled final events was developed leading to the completion of the natural product.
Introduction
Biologically active molecules isolated from marine organisms
constitute an ever-growing subset of all natural products
collected, and among them are some of the most potent
antitumor and cytotoxic agents yet discovered.1 The dragma-
cidins (Figure 1, 1-7) represent an emerging class of bioactive
marine natural products obtained by an exhaustive set of
protocols from a number of deep water sponges including
Dragmacidon, Halicortex, Spongosorites, and Hexadella, and
the tunicate Didemnum candidum.2,3 The initial four dragma-
cidins identified contained a piperazine linker (Figure 1, 1-4)
and displayed modest antifungal, antiviral, and cytotoxic activi-
ties. Recently, the structurally more complex aminoimidazole-
and guanidine-containing pyrazinone dragmacidins D (5), E (6),
and F (7) were isolated and shown to possess a wide range of
interesting biological properties as well.
In particular, dragmacidin D (5) is a potent inhibitor of serine-
threonine protein phosphatases (PP), and preliminary evidence
suggests that dragmacidin D is a selective inhibitor of PP1 versus
PP2A.3b It is also an in vivo nonsteroidal antiinflammatory agent
as shown by its potent inhibition of resiniferitoxin-induced
inflammation in mouse ear models.4 Additionally, dragmacidin
D was found to selectively inhibit neural nitric oxide synthase
(bNOS) in the presence of inducible NOS (iNOS).5 Endogenous
nitric oxide (NO), produced via NOS-mediated metabolism of
Figure 1. The dragmacidin alkaloids.
L-arginine, is known to play a role in a variety of regulatory
functions such as the control of blood pressure, antibacterial
activity, gastric motility, and neurotransmission.6 Alternatively,
the overproduction of nitric oxide by certain NOS isozymes
has been implicated in a host of inflammatory diseases, such
as arthritis7 and asthma,8 as well as some neurodegenerative
disorders. Therefore, the ability to selectively inhibit the brain
NOS isozyme (bNOS) may be useful in a variety of therapeutic
areas including the treatment of Alzheimer’s, Parkinson’s, and
Huntington’s diseases.9
* To whom correspondence should be addressed. E-mail: stoltz@
caltech.edu.
(1) (a) Norcross, R. D.; Paterson, I. Chem. ReV. 1995, 95, 2041-2114. (b)
Faulkner, D. J. Nat. Prod. Rep. 2002, 19, 1-48.
(2) (a) Kohmoto, S.; Kashman, Y.; McConnell, O. J.; Rinehart, K. L., Jr.;
Wright, A.; Koehn, F. J. Org. Chem. 1988, 53, 3116-3118. (b) Morris, S.
A.; Andersen, R. J. Tetrahedron 1990, 46, 715-720. (c) Fahy, E.; Potts,
B. C. M.; Faulkner, D. J.; Smith, K. J. Nat. Prod. 1991, 54, 564-569.
(3) (a) Wright, A. E.; Pomponi, S. A.; Cross, S. S.; McCarthy, P. J. Org. Chem.
1992, 57, 4772-4775. (b) Capon, R. J.; Rooney, F.; Murray, L. M.; Collins,
E.; Sim, A. T. R.; Rostas, J. A. P.; Butler, M. S.; Carroll, A. R. J. Nat.
Prod. 1998, 61, 660-662. (c) Cutignano, A.; Bifulco, G.; Bruno, I.;
Casapullo, A.; Gomez-Paloma, L.; Riccio, R. Tetrahedron 2000, 56, 3743-
3748.
(6) (a) Fukuto, J. M.; Chaudhuri, G. Annu. ReV. Pharmacol. Toxicol. 1995,
35, 165-194. (b) Yun, H. Y.; Dawson, V. L.; Dawson, T. M. Crit. ReV.
Neurobiol. 1996, 10, 291-316.
(7) Amin, A. R.; Attur, M.; Abramson, S. B. Curr. Opin. Rheumatol. 1999,
11, 202-209.
(8) (a) Baraldi, E.; Dario, C.; Ongaro, R.; Scollo, M. Am. J. Respir. Crit. Care
Med. 1999, 159, 1284-1288. (b) Sanders, S. P. Proc. Soc. Exp. Biol. Med.
1999, 220, 123-132. (c) Stirling, R. G.; Kharitonov, S. A.; Campbell, D.;
Robinson, D. S.; Durham, S. R.; Chung, K. F.; Barnes, P. J. Thorax 1998,
53, 1030-1034.
(9) (a) Molina, J. A.; Jimenez-Jimenez, F. J.; Orti-Pareja, M.; Navarro, J. A.
Drugs Aging 1998, 12, 251-259. (b) Thorns, V.; Hansen, L.; Masliah, E.
Exp. Neurol. 1998, 150, 14-20.
(4) Jacobs, R. S.; Pomponi, S.; Gunasekera, S.; Wright, A. PCT Int. Appl.
WO 9818466, May 7, 1998.
(5) Longley, R. E.; Isbrucker, R. A.; Wright, A. E. U.S. Patent 6,087,363,
July 7, 2000.
9
10.1021/ja027822b CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 13179-13184
13179

A R T I C L E S
Garg et al.
Scheme 1. Dragmacidin D Retrosynthetic Plan
Figure 2. The potential biosynthetic relationship of dragmacidins D, E,
and F.
Given the biological relevance of the dragmacidins, especially
the pyrazinone-containing dragmacidins D, E, and F, it is
surprising that relatively little synthetic chemistry has been
published regarding these structures.10 Moreover, due to the
general difficulty in isolating the dragmacidins, nature is an
impractical source for obtaining the large quantities of material
necessary for advanced biological testing and analogue synthesis.
Herein, we outline an efficient synthesis of the common core
structure of dragmacidins D, E, and F, which we have prepared
by the implementation of sequential temperature-modulated
Suzuki cross-coupling reactions that are exquisitely selective.
Additionally, we have elaborated this core by a delicate series
of highly specific transformations to accomplish the first total
synthesis of dragmacidin D (5).
Scheme 2. Alternate Dragmacidin D Retrosynthesis
Results and Discussion
Synthetic Planning. Our synthetic planning in the area of
the dragmacidins began with the recognition that the complex
dragmacidins D, E, and F are likely biosynthetically related.11
Of the possible biosynthetic scenarios, most probable is the
notion that dragmacidins E and F are derived by cyclization of
either dragmacidin D or a closely related congener (Figure 2).
For example, Friedel-Crafts type cyclization could result in
the conversion of D to E (i.e., 5 f 6), while an oxidative
dearomatization with concomitant cyclization could facilitate
the formation of the unique polycyclic framework present in
dragmacidin F (i.e., 5 f 7). Therefore, the identification of
dragmacidin D as a suitable biosynthetic precursor to dragma-
cidins E and F led to its designation by us as a reasonable first
synthetic target in this class of complex alkaloids.
for example: (A) the seven nitrogen atoms present in the
compact structure, (B) the four differentially substituted het-
erocycles, (C) the nontrivial nature of the substitution patterns
on the 3,6-disubstituted and 3,4,7-trisubstituted indoles, and (D)
the highly polar and reactive nature of the aminoimidazolium
subunit and of potential precursors to the natural product.
Presented in Scheme 1 is our retrosynthetic strategy for the
synthesis of dragmacidin D. As a key strategic maneuver, we
chose to introduce the aminoimidazolium functionality of
dragmacidin D at a late stage in the synthesis to facilitate the
handling of key intermediates. Thus, disconnection of the natural
product (5) provides silyl ether 8 (Scheme 1), which was
envisioned to arise by a stepwise, palladium-catalyzed three-
component coupling of fragments 9 + 10 + 11. Critical to the
success of this plan was the capacity to utilize metallo-
bromoindole 11 directly in the cross-coupling sequence, thereby
defining a strategic challenge that would limit the options
available for protecting groups as well as end-game scenarios.
The indole building blocks 12 and 13 were readily available
from simple aromatic starting materials 14 and 15.
Although the structure of dragmacidin D appears deceptively
simple, closer consideration reveals several synthetic challenges,
(10) Although some elegant synthetic work directed at the piperazine-containing
dragmacidins has appeared, reports directed toward those members pos-
sessing guanidine-functionalized indoles around a pyrazinone core have
been limited. For synthetic work aimed at the piperazine-containing
dragmacidins, see: (a) Jiang, B.; Smallheer, J. M.; Amaral-Ly, C.; Wuonola,
M. A. J. Org. Chem. 1994, 59, 6823-6827. (b) Whitlock, C. R.; Cava, M.
P. Tetrahedron Lett. 1994, 35, 371-374. (c) Kawasaki, T.; Enoki, H.;
Matsumura, K.; Ohyama, M.; Inagawa, M.; Sakamoto, M. Org. Lett. 2000,
2, 3027-3029. (d) Miyake, F. Y.; Yakushijin, K.; Horne, D. A. Org. Lett.
2000, 2, 3185-3187. (e) Yang, C.-G.; Wang, J.; Tang, X.-X.; Jiang, B.
Tetrahedron: Asymmetry 2002, 13, 383-394. (f) Kawasaki, T.; Ohno, K.;
Enoki, H.; Umemoto, Y.; Sakamoto, M. Tetrahedron Lett. 2002, 43, 4245-
4248. For studies specifically targeting dragmacidins D, E, or F, see: (g)
Jiang, B.; Gu, X.-H. Bioorg. Med. Chem. 2000, 8, 363-371. (h) Jiang, B.;
Gu, X.-H. Heterocycles 2000, 53, 1559-1568. (i) Yakushijin, K.; Horne,
D. A. PCT Int. Appl. WO 0194310 A1, December 13, 2001. (j) Yang,
C.-G.; Wang, J.; Jiang, B. Tetrahedron Lett. 2002, 43, 1063-1066. (k)
Miyake, F. Y.; Yakushijin, K.; Horne, D. A. Org. Lett. 2002, 4, 941-943.
(11) For a brief discussion, see refs 3b and 3c.
An Initial Cyclocondensation Digression. In addition to the
strategy described above, we considered an alternative method
to construct the bis-indole framework wherein disconnection
of the pyrazinone 16 directly revealed two fragments, keto
aldehyde 17 and aminoamide 18 (Scheme 2). These function-
alized indoles were likewise available from the parent indoles
12 and 13. Aside from the obvious advantage of increased
convergence, this approach avoided selectivity issues with regard
9
13180 J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002

The First Total Synthesis of Dragmacidin D
A R T I C L E S
Scheme 4. Preparation of the 3,4,7-Trifunctionalized Indole
Scheme 3. A Cyclocondensative Pyrazinone Synthesis
Subunit
to halide/organometallic couplings alluded to in the transition-
metal approach described above.
reaction (Scheme 4).15 Thus, treatment of nitroaromatic 14 with
3.5 equiv of vinyl Grignard reagent produced the benzyloxy
indole 12a directly in modest yield. Following initial protection
of the indole nitrogen by a 2-(trimethylsilyl)ethoxymethyl (SEM)
group,16 halogen-metal exchange and trapping by dioxaboro-
lane reagent 25 produced metalloindole 26.17 Suzuki coupling
of indole 26 and vinyl bromide 2718 provided ether 28.19 Final
conversion of ether 28 to coupling fragment 9 was accomplished
using a sequence involving selective hydrogenation of the
terminal olefin,20 bromination at the C(3) position,21 and
halogen-metal exchange/trapping with dioxaborolane 25 (66%
yield, three steps).
The differentially substituted pyrazine core structure (e.g.,
10, Scheme 5) was readily available by iodination of known
aminopyrazine 2922 via the in situ prepared diazonium salt.23
Preparation of the bromoindole boronic acid derivative 11 from
the parent 1324 proceeded by protection of the indole nitrogen,25
The appeal of the cyclocondensation approach was heightened
by the success of a simple model system that constituted one
of our first experiments in this area. Synthesis of the unfunc-
tionalized keto aldehyde 20 and aminoamide 21 was straight-
forward from the well-known product of the reaction of indole
with oxalyl chloride (i.e., 19).12 Modified Rosenmund reduction
of acid chloride 19 by the action of Bu3SnH produced aldehyde
20 in modest yield.13 Alternatively, treatment of 19 with
ammonia, followed by oximation of the resulting ketoamide,
and reduction with H2 over Pd/C in MeOH produced aminoa-
mide 21. In the key cyclocondensation experiment, simply
heating keto aldehyde 20 and aminoamide 21 in aq KOH at 70
°C led to the formation of pyrazinone 22 in 75% yield (Scheme
3).
With this result in hand, we investigated the synthesis and
cyclocondensation of a number of highly functionalized systems
more closely aligned with the target molecule (i.e., 5). Our
enthusiasm for this direct cyclocondensative coupling approach
was quickly diminished, however, as all attempts to synthesize
more advanced pyrazinones produced none of the desired
dragmacidin framework. For example, attempted coupling of
keto aldehydes 23a-c with aminoamide 18 under a variety of
basic and acidic conditions provided only traces (if any) of the
pyrazinone products (eq 1). It became clear that cycloconden-
sation approaches to the dragmacidins that involved substitution
at the C(4) position of indoles of the type 23 would not be
feasible in our hands.14
(14) Although there is clearly an electronic difference between indoles 20, 21
and 23a-c, 18, respectively, the steric component of C(4) substitution
appears to be the determining factor as illustrated by the reaction of i and
18 to produce the desired pyrazinone ii in good yield.
(15) (a) Bartoli, G.; Palmieri, G.; Bosco, M.; Dalpozzo, R. Tetrahedron Lett.
1989, 30, 2129-2132. (b) Bosco, M.; Dalpozzo, R.; Bartoli, G.; Palmieri,
G.; Petrini, M. J. Chem. Soc., Perkin Trans. 2 1991, 5, 657-663.
(16) (a) Edwards, M. P.; Ley, S. V.; Lister, S. G.; Palmer, B. D. J. Chem. Soc.,
Chem. Commun. 1983, 630-633. (b) Piers, E.; Britton, R.; Andersen, R.
J. J. Org. Chem. 2000, 65, 530-535.
(17) Belletete, M.; Beaupre, S.; Bouchard, J.; Blondin, P.; Leclerc, M.; Durocher,
G. J. Phys. Chem. B 2000, 104, 9118-9125.
(18) Li, K.; Du, W.; Que, N. L. S.; Liu, H.-W. J. Am. Chem. Soc. 1996, 118,
8763-8764.
(19) For general reviews on Suzuki couplings, see: (a) Miyaura, N.; Suzuki,
A. Chem. ReV. 1995, 95, 2457-2483. (b) Suzuki, A. J. Organomet. Chem.
1999, 576, 147-168.
(20) (a) Maki, S.; Okawa, M.; Matsui, R.; Hirano, T.; Niwa, H. Synlett 2001,
10, 1590-1592. (b) In the presence of minor impurities and/or solutions
of benzene that were not presaturated with H2, the reaction produced a
major byproduct identified as enol ether iii.
Synthesis of the Building Blocks. With the difficulties of
the cyclocondensation approach clearly illuminated, we inves-
tigated our initial palladium coupling approach to dragmacidin
D. As a starting point for the synthesis, the parent 4,7-
disubstituted indole was accessed by employing the Bartoli
(21) (a) Ayer, W. A.; Craw, P. A.; Ma, Y.-T.; Miao, S. Tetrahedron 1992, 48,
2919-2924. (b) Amat, M.; Hadida, S.; Sathyanarayana, S.; Bosch, J. J.
Org. Chem. 1994, 59, 10-11.
(12) (a) Kharasch, M. S.; Kane, S. S.; Brown, H. C. J. Am. Chem. Soc. 1940,
62, 2242-2243. (b) Hashem, M. A.; Sultana, I.; Hai, M. A. Indian J. Chem.,
Sect. B 1999, 38, 789-794.
(13) (a) Vereshchagin, A. L.; Bryanskii, O. V.; Semenov, A. A. Chem.
Heterocycl. Compd. (Engl. Transl.) 1983, 19, 40-42. (b) Kuivila, H. G. J.
Org. Chem. 1960, 25, 284-285.
(22) Barlin, G. B. Aust. J. Chem. 1983, 36, 983-992.
(23) Zhu, Z.; Moore, J. S. J. Org. Chem. 2000, 65, 116-123.
(24) Prepared from 15 by a modified Leimgruber-Batcho synthesis, see:
Schumacher, R. W.; Davidson, B. S. Tetrahedron 1999, 55, 935-942.
9
J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002 13181

A R T I C L E S
Garg et al.
Scheme 5. Synthesis of the Pyrazine and Bromoindole
Fragments
treatment with Hg(OAc)2, and reaction of the resulting orga-
nomercurial 30 with BH3‚THF followed by hydrolytic workup
(Scheme 5, 82% overall yield from 13).26
Fragment Coupling. Having efficient means to access the
desired building blocks for the preparation of the dragmacidin
D skeleton in hand, we turned our attention toward experiments
that would delineate suitable conditions for coupling. We
initially surveyed a variety of coupling reactions involving model
indoles and various halogenated pyrazine derivatives to gauge
the relative reactivity of such systems as well as the suitability
of the protective groups on the indole nitrogen. It was quickly
established that halogenated pyrazines are highly reactive toward
palladium-mediated couplings to metalated indoles. Furthermore,
particularly useful for the proposed synthesis of dragmacidin
D is the fact that the oxidative addition of palladium(0) to
pyrazinyl halides is more facile than that to simple aromatic
halides, as has been established in the literature.27 For example,
coupling of borylated indoles 31a and 31b with readily available
chloropyrazine 3228 proceeded smoothly at 80 °C under standard
Suzuki conditions (eq 2). Under identical conditions, simple aryl
chlorides do not readily participate in such couplings.29 Ad-
ditionally, treatment of chloroiodopyrazine 34 with 2 equiv of
indole 31a at 23 °C produced indolopyrazine 35a exclusively,
while raising the temperature to 80 °C resulted in the formation
of the bis-indolopyrazine 36 (eq 3). Again these results were in
good accord with expectations derived from literature observa-
tions. A more surprising development was observed upon
treatment of pyrazine 34 with an excess of silylated boronic
ester 31b (2.3 equiv) at 80 °C. Under these conditions,
exclusively monocoupled product was obtained as a mixture
of silylated and desilylated compounds (eq 4, 35b and 35c).
Clearly, this dichotomy points to a remote electronic effect of
the indole protecting group for the activation of the intermediate
chloroindolopyrazine (35) toward coupling. Thus, our choice
of a tosyl protective group for the bromoindole fragment proved
fortuitous. Furthermore, to maximize position selectivity in the
couplings leading to dragmacidin D, chloroiodopyrazine 34 gave
way to bromoiodopyrazine 10 as the optimal synthetic linchpin.
With regard to the union of building blocks 9, 10, and 11,
we have found that room-temperature Suzuki coupling of
dihalopyrazine 10 and indole 11 proceeds selectively to the
coupled indolopyrazine 37 (see Scheme 6). In the critical second
Suzuki coupling of dibromide 37 with metallo-indole 9, we were
delighted to find that under carefully controlled conditions (50
°C, 72 h), the desired bis-indole alkoxy pyrazine 8 was formed
in good yield and with complete selectivity for coupling of the
pyrazinyl bromide in the presence of the indolyl bromide. As
predicted from the model studies above (eqs 2-4), precise
temperature control is needed for this sequence of coupling
reactions, particularly in the reaction of 9 + 37 f 8. At
temperatures approaching 80 °C, coupling of the bromoindole
unit becomes competitive with the desired mode of coupling.
Thus, it was particularly important to perform the reaction at
the lowest possible temperature (50 °C), to maximize selectivity
in this sequence.
Scheme 6. Sequential Selective Suzuki Couplings
End-Game Strategy 1. With the suitably elaborated bis-
indolopyrazine core structure available from the Suzuki chem-
istry, we initiated our final approach to dragmacidin D. Selective
cleavage of the silyl ether in 8 was accomplished by the action
of HF‚pyridine,30 and the resulting primary alcohol was oxidized
to carboxylic acid 38a (Scheme 7). Arndt-Eistert homologation
to R-bromo ketone 39 was initially problematic and furnished
essentially none of the desired product (ca. 10% yield). Instead,
(25) Wenkert, E.; Moeller, P. D. R.; Piettre, S. R.; McPhail, A. T. J. Org. Chem.
1988, 53, 3170-3178.
(26) Zheng, Q.; Yang, Y.; Martin, A. R. Heterocycles 1994, 37, 1761-1772.
(27) For an excellent discussion of the cross-coupling chemistry of pyrazines,
see: Li, J. J.; Gribble, G. W. Palladium in Heterocyclic Chemistry:
A
Guide for the Synthetic Chemist; Pergamon: Amsterdam, 2000; pp 355-
373.
(28) Turck, A.; Ple´, N.; Dognon, D.; Harmoy, C.; Que´guiner, G. J. Heterocycl.
Chem. 1994, 31, 1449-1454.
(29) For a discussion, see: Grushin, V. V.; Alper, H. Chem. ReV. 1994, 94,
1047-1062.
(30) For the cleavage of silyl ethers with HF‚pyridine, see: Nicolaou, K. C.;
Webber, S. E. Synthesis 1986, 6, 453-461.
9
13182 J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002

The First Total Synthesis of Dragmacidin D
A R T I C L E S
Scheme 7. Studies toward the Synthesis of Dragmacidin D
Scheme 8. The Synthesis of Dragmacidin D
nearly a quantitative recovery of the corresponding methyl ester
38b was obtained, pointing to rapid hydrolysis of the intermedi-
ate acid chloride and subsequent esterification by diazomethane.
This result was particularly surprising given that model studies
performed on analogous indolepropionic acid derivatives lacking
substitution at the indole C(3) position proceeded without event.
However, after careful and rigorous drying of the diazomethane
over KOH followed by sodium metal,31 we were able to obtain
satisfactory yields of bromoketone 39 (58% yield). Again in
stark contrast to model studies performed on simple derivatives,
treatment of bromoketone 39 with acetylguanidine produced
none of the desired aminoimidazole derivative 40; rather, nearly
a quantitative yield of acetoxyketone 41 was recovered.
Recognizing that substitution chemistry at the R-position of
ketone 39 must occur to produce acetate 41, we treated
bromoketone 39 with a saturated NH3/MeOH solution to
produce an aminoketone, which was converted to aminoimi-
dazole 42 upon exposure to cyanamide in EtOH at 70 °C.32
Although aminoimidazole 42 represented a protected form of
the natural product 5, all attempts to remove the four remaining
protective groups resulted in nonspecific cleavage of the
aminoimidazole moiety. Thus, compound 42 symbolized an
unsatisfying dead-end for our synthesis.
to aldehyde 43 using the Dess-Martin periodinane reagent
(Scheme 8).33,34 As an alternative to the problematic Arndt-
Eistert homologation sequence discussed above, nitromethane
addition35 and subsequent oxidation to nitroketone 44 proceeded
smoothly (98% yield, two steps). Importantly, this protocol
provided a molecule (44) possessing the desired homologous
carbon framework and delivered a nitrogen atom to the
R-position of the ketone in two high yielding steps. With
nitroketone 44 in hand, the cleavage of both indole nitrogen
protective groups was readily accomplished. Scrupulously
deoxygenated ethanolic KOH facilitated the removal of the
N-tosyl group,36 and LiBF4 followed by aqueous NaOH effected
complete hydrolysis of the SEM (i.e., 44 f 45). Selective
reduction of nitroketone 45 using stannous chloride37 and
removal of the benzyl and methyl ethers with iodotrimethylsilane
revealed the fully deprotected aminoketone 46.38 Particularly
noteworthy is that this sequence includes reduction of the nitro
group to an amine and cleavage of the benzyl ether unit without
the need for catalytic hydrogenative procedures or other methods
that would have resulted in debromination of the indole C(6)
position. Final installation of the aminoimidazolium unit with
(33) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155-4156.
(34) (a) Aldehyde 43 was the last compound in the synthesis that was readily
purified on preparative scale by silica gel flash chromatography. Although
analytical scale purification beyond this point was enabled by reversed-
phase HPLC or thin-layer chromatography on SiO2, all preparative scale
purification was conducted by reversed-phase C18 chromatography. Dif-
ficulties associated with the separation of similarly polar compounds by
this method necessitated that all reactions in the sequence leading from 43
f 5 be high yielding. (b) The highly fluorescent nature (λexc ) 365 nm)
of many intermediates along the pathway from 8 f 5 facilitated the isolation
of small amounts of compound in large quantities of solvent (i.e., ca. 1
mg/30-50 mL of solvent during reversed-phase chromatography). See the
Supporting Information for details.
Completion of the Total Synthesis. Unabated by our
inability to transform 42 into 5, we focused our attention on an
end-game scenario that included the earlier deprotection of
masked functionality and postponement of the aminoimidazole
introduction. We also sought an alternative homologation
protocol, as the Arndt-Eistert method had become our synthetic
bottleneck. After extensive experimentation, our second-genera-
tion sequence was initiated by selective silyl ether cleavage of
Suzuki adduct 8 and oxidation of the resulting primary alcohol
(35) (a) Luzzio, F. A. Tetrahedron 2001, 57, 915-945. (b) Ayerbe, M.; Arrieta,
A.; Coss´ıo, F. P.; Linden, A. J. Org. Chem. 1998, 63, 1795-1805.
(36) Gilbert, E. J.; Chisholm, J. D.; Van Vranken, D. L. J. Org. Chem. 1999,
64, 5670-5676.
(31) Arndt, F. Org. Synth. 1943, Coll. Vol. 2, 165.
(32) (a) Lancini, C. C.; Lazzari, E.; Sartori, C. J. Antibiot. 1968, 21, 387-392.
(b) Howes, P. D.; Cleasby, A.; Evans, D. N.; Feilden, H.; Smith, P. W.;
Sollis, S. L.; Taylor, N.; Wonacott, A. J. Eur. J. Med. Chem. 1999, 34,
225-234.
(37) Sakowski, J.; Bo¨hm, M.; Sattler, I.; Dahse, H.-M.; Schlitzer, M. J. Med.
Chem. 2001, 44, 2886-2899.
(38) Lott, R. S.; Chauhan, V. S.; Stammer, C. H. J. Chem. Soc., Chem. Commun.
1979, 495-496.
9
J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002 13183

A R T I C L E S
Garg et al.
cyanamide followed by treatment with trifluoroacetic acid
provided dragmacidin D (5) in 86% yield. Synthetic dragmacidin
D was spectroscopically identical to samples obtained from
natural sources (1H NMR, 13C NMR, IR, HRMS, UV).
Interestingly, the exact order of final synthetic events
presented herein was essential for the completion of dragmacidin
D. In particular, intermediates 44-46 were highly labile when
treated under a variety of other conditions. For example, attempts
to reduce nitroketone 44 or to remove the SEM group prior to
detosylation resulted in substantial nonspecific decomposition.
Likewise, efforts to deprotect 45 prior to nitro reduction led to
the decomposition of the nitroketone moiety. Finally, reversing
the order of the final two steps (i.e., aminoimidazole formation
followed by treatment with TMSI) afforded only a low yield
of dragmacidin D (ca. 5%).
to dragmacidin D, and the extension of this chemistry toward
the synthesis of dragmacidins E (6) and F (7) by both biomimetic
and de novo synthetic routes. Results in these areas will be
presented in due course.
Acknowledgment. The authors are grateful to the California
Institute of Technology, the Camille and Henry Dreyfus
Foundation (New Faculty Award to B.M.S.), Merck Research
Laboratories, NIH-NIGMS (R01 GM65961-01), DOD (NDSEG
graduate fellowship to N.K.G.), and Pfizer‚UNCF (postdoctoral
fellowship to R.S.) for generous financial support. The Dervan
and MacMillan labs are acknowledged for helpful discussions
and the generous use of instrumentation. Michael Marques, Wei
Zhang, and Yen Nguyen are acknowledged for experimental
assistance. We thank John Greaves (University of California,
Irvine) for obtaining high-resolution mass spectra and Scott Ross
of the Beckman Institute (Caltech) for assistance in obtaining
NMR spectra. We gratefully acknowledge Dr. Amy E. Wright
of the Harbor Branch Oceanographic Institute (North Fort Pierce,
Florida) and Professor Robert J. Capon of the University of
Melbourne (Parkville, Victoria, Australia) for kindly providing
samples and spectra of natural dragmacidin D.
Conclusion
In summary, we have completed the first total synthesis of
the important bis-indole alkaloid dragmacidin D (5). The concise
route that we have developed (longest linear sequence of 17
steps from 14) relies on a key series of temperature-controlled
selective Suzuki couplings and a meticulous sequence of final
events leading to the completion of the natural product. Of
particular note are the unique reactivity of 3,4,7-trisubstituted
indoles and the steric and electronic subtleties associated with
these systems. Specifically, the interplay of the indole C(3)-
C(4) positions was observed in both the cyclocondensation
approach and the end-game approaches to dragmacidin D. We
are currently investigating the synthesis of a number of
dragmacidin analogues, the development of an asymmetric route
Supporting Information Available: Full experimental details
and characterization data for all synthetic intermediates, and
spectral comparison for synthetic and natural 5 (PDF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA027822B
9
13184 J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002