Total synthesis and initial structure - Function analysis of the potent V-ATPase inhibitors salicylihalamide A and related compounds

Article abstract of DOI:10.1021/ja0177713

Salicylihalamide A is the first member of a growing class of macrocyclic salicylate natural products that induce a variety of interesting phenotypes in cultured mammalian cells. Salicylihalamide A was reported to be a unique and highly differential cytotoxin and a potent inhibitor of the mammalian vacuolar (H⁺)ATPase. The total synthesis of both enantiomers of salicylihalamide A, a revision of the absolute configuration assigned to the natural product, and extensive structure-function studies with synthetic salicylihalamide variants are reported. These studies were possible only due to a highly efficient synthetic strategy that features (1) a remarkably E-selective ring-closing olefin metathesis to construct the 12-membered benzolactone skeleton 29, (2) a mild stereocontrolled elaboration to E-alkenyl isocyanate 41, and (3) addition of carbon, oxygen, and sulfur nucleophiles to isocyanate 41 to obtain salicylihalamide A and congeners. We demonstrate for the first time that salicylihalamide A is a potent inhibitor of fully purified reconstituted V-ATPase from bovine brain, and have identified several similarly potent side chain modified derivatives, including salicylihalamide dimers 43-45. In combination, these studies have laid the foundation for ongoing studies aimed at a comprehensive understanding of salicylihalamide's mode-of-action, of potential relevance to the development of lead compounds for the treatment of osteoporosis and cancer.

Full text of DOI:10.1021/ja0177713

Published on Web 03/06/2002
Total Synthesis and Initial Structure-Function Analysis of the
Potent V-ATPase Inhibitors Salicylihalamide A and Related
Compounds
Yusheng Wu,† Xibin Liao,† Ruifang Wang,† Xiao-Song Xie,‡ and
Jef K. De Brabander*,†
Contribution from the Departments of Biochemistry and Internal Medicine,
The UniVersity of Texas Southwestern Medical Center at Dallas, 5323 Harry Hines BouleVard,
Dallas, Texas 75390-9038
Received December 14, 2001
Abstract: Salicylihalamide A is the first member of a growing class of macrocyclic salicylate natural products
that induce a variety of interesting phenotypes in cultured mammalian cells. Salicylihalamide A was reported
to be a unique and highly differential cytotoxin and a potent inhibitor of the mammalian vacuolar (H+)-
ATPase. The total synthesis of both enantiomers of salicylihalamide A, a revision of the absolute configuration
assigned to the natural product, and extensive structure-function studies with synthetic salicylihalamide
variants are reported. These studies were possible only due to a highly efficient synthetic strategy that
features (1) a remarkably E-selective ring-closing olefin metathesis to construct the 12-membered
benzolactone skeleton 29, (2) a mild stereocontrolled elaboration to E-alkenyl isocyanate 41, and (3) addition
of carbon, oxygen, and sulfur nucleophiles to isocyanate 41 to obtain salicylihalamide A and congeners.
We demonstrate for the first time that salicylihalamide A is a potent inhibitor of fully purified reconstituted
V-ATPase from bovine brain, and have identified several similarly potent side chain modified derivatives,
including salicylihalamide dimers 43-45. In combination, these studies have laid the foundation for ongoing
studies aimed at a comprehensive understanding of salicylihalamide’s mode-of-action, of potential relevance
to the development of lead compounds for the treatment of osteoporosis and cancer.
Introduction
cytotoxicity profile in the NCI 60-cell line human tumor assay
(mean panel GI50’s: 1.6-15 nM; range of differential sensitivity
During the past decades, products of secondary metabolic
pathways have fueled the pharmaceutical research enterprise
and, for example, have provided valuable lead compounds for
the treatment of various human diseases. Furthermore, small
molecule ligands that elicit specific and unique biological
responses in mammalian cells also constitute extremely valuable
research tools in discovery biology efforts. In this context, a
unique opportunity is provided by the discovery of a variety of
unusual macrocyclic salicylate natural products that were
isolated from both terrestrial and marine sources based on their
ability to induce a particular phenotype in mammalian cells.
In 1997, scientists at the Laboratory of Drug Discovery
Research and Development at the National Cancer Institute
reported the bioassay-guided isolation of salicylihalamides A
and B (1a,b) from an unidentified species of the marine sponge
Haliclona (Figure 1).1 Shortly afterward, the same group
identified a family of related metabolites, termed lobatamides
A-F, in three different tunicate species of the genus Aplidium.2
Notably, these natural products exhibit a unique differential
g103). Subsequent years witnessed the isolation of structurally
related compounds from a variety of terrestrial microorganisms.
Apicularen A was isolated from a myxobacteria based on its
extremely potent cytostatic activity against human cancer cell
lines including a multi-drug-resistant cervix carcinoma cell line.3
The same report mentions anecdotally that apicularen A induces
several abnormal effects including the formation of mitotic
spindles with multiple spindle poles and clusters of bundled
actin from the cytoskeleton and the induction of an apoptotic-
like cell death. Pfizer Pharmaceuticals isolated the fungal
metabolites CJ-12,950 and CJ-13,357 based on their ability to
induce Low-Density Lipoprotein (LDL) receptor gene expres-
sion, an observation that has potential relevance to the treatment
of hypercholesterolemia and hyperlipidemia.4 Oximidines I and
II are the latest additions to this growing class of enamide-
(2) (a) McKee, T. C.; Galinis, D. L.; Pannell, L. K.; Cardellina, J. H., II; Laakso,
J.; Ireland, C. M.; Murray, L.; Capon, R. J.; Boyd, M. R. J. Org. Chem.
1998, 63, 7805-7810. (b) Lobatamide A is identical to the structure of
YM-75518, see: Suzumura, K.-I.; Takahashi, I.; Matsumoto, H.; Nagai,
K.; Setiawan, B.; Rantiatmodjo, R. M.; Suzuki, K.-I.; Nagano, N.
Tetrahedron Lett. 1997, 38, 7573-7576.
(3) (a) Kunze, B.; Jansen, R.; Sasse, F.; Ho¨fle, G.; Reichenbach, H. J. Antibiot.
1998, 51, 1075-1080. (b) Jansen, R.; Kunze, B.; Reichenbach, H.; Ho¨fle,
G. Eur. J. Org. Chem. 2000, 913-919.
(4) Dekker, K. A.; Aiello, R. J.; Hirai, H.; Inagaki, T.; Sakakibara, T.; Suzuki,
Y.; Thompson, J. F.; Yamauchi, Y.; Kojima, N. J. Antibiot. 1998, 51, 14-
20.
* To whom correspondence should be addressed. Phone: 214-648-7808.
Fax: 214-648-6455. E-mail: jdebra@biochem.swmed.edu.
† Department of Biochemistry.
‡ Department of Internal Medicine.
(1) Erickson, K. L.; Beutler, J. A.; Cardellina, J. H., II; Boyd, M. R. J. Org.
Chem. 1997, 62, 8188-8192; correction: J. Org. Chem. 2001, 66, 1532.
9
10.1021/ja0177713 CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 3245-3253
3245

A R T I C L E S
Wu et al.
Scheme 1. Installation of the N-Acyl Enamine Side Chain
well as derived probe reagents. Herein, we present a full account
of our salicylihalamide synthetic venture. Initiated in the context
of generating informative structure-function data, these efforts
have culminated in a highly efficient total synthesis of sali-
cylihalamide A and modified congeners.9-11
Architecture and Synthetic Strategy
The structure and relative stereochemistry of salicylihalamide
A was assigned based on extensive data gathered from a variety
of high-field NMR experiments.1 The absolute configuration
of (-)-salicylihalamide A, initially assigned through Mosher
ester H NMR experiments,1 was later revised by us to 12S,-
1
Figure 1. Salicylihalamide A and Structurally Related Natural Products.
13R,15S through total synthesis.9,12 Salicylihalamides as well
as the biosynthetically related macrocyclic salicylates represent
a structurally novel class of macrocyclic benzolactones incor-
porating salicylic acid, and are decorated with an unusual, highly
unsaturated, N-acyl enamine side chain. Given the sensitive
nature of this common construct, we deemed it crucial to
introduce this at a late stage in the synthesis utilizing a unifying
and mild strategy. Considering the options, we felt that the
addition of 1-metallo-1,3-hexadiene (I, X-Y ) CH-CH2) or
the corresponding iminoalkenyl derivative (I, X-Y ) N-O)
to a stereodefined E-alkenyl isocyanate II would offer the
distinct advantage of mild reaction conditions and control of
stereochemistry (Scheme 1).13,14 Isocyanate II is to be derived
from the corresponding R,â-unsaturated carboxylic acid III (acyl
azide formation/Curtius rearrangement), in turn accessible from
alcohol IV via oxidation/Horner-Wadsworth-Emmons ho-
mologation.
substituted cyclic salicylates and were shown to induce selective
growth inhibition of oncogene-transformed rat fibroblasts (3Y1
cells) at 15- to 30-fold lower concentrations than for the parent
cell line.5
Salicylihalamides evoked particular interest due to the
provocative observation that pattern-recognition analysis (COM-
PARE) of their activity profiles in the NCI 60-cell line screen
initially did not reveal any significant correlation to the profiles
acquired for other antitumor compounds contained within the
NCI standard agent database.1 Such unique signature profiles
are indicative of a potentially novel mechanism of antineoplastic
activity. Using a more extensive database analysis, the NCI team
subsequently found that the 60-cell profiles of salicylihalamides,
lobatamides, and oximidines gave consistently high correlations
with the historical database profiles of bafilomycins, prototypical
vacuolar (H+)-ATPase (V-ATPase) inhibitors.6 A common or
similar mode-of-action with bafilomycins was further suggested
based on in vitro studies with crude membrane preparations of
mammalian V-ATPases.6 Importantly, these studies have re-
vealed that, unlike bafilomycins, salicylihalamide A exquisitely
discriminates between mammalian and nonmammalian V-
ATPases. The implication of aberrant V-ATPase function in
many different human diseases, such as osteoporosis and cancer,
warrants the development of synthetically tractable and biologi-
cally selective pharmacological modulators of V-ATPases.7,8 To
study the molecular pharmacology of this unique class of
macrocyclic salicylates, we initiated a synthetic program aimed
at producing sufficient quantities of these natural products as
(9) Part of this work has appeared as preliminary communications: (a) Wu,
Y.; Esser, L.; De Brabander, J. K. Angew. Chem., Int. Ed. 2000, 39, 4308-
4310. (b) Wu, Y.; Seguil, O. R.; De Brabander, J. K. Org. Lett. 2000, 2,
4241-4244.
(10) Subsequent total syntheses have appeared: (a) Snider, B. B.; Song, F. Org.
Lett. 2001, 3, 1817-1820. (b) Smith, A. B., III; Zheng, J. Synlett 2001,
1019-1023. (c) Labrecque, D.; Charron, S.; Rej, R.; Blais, C.; Lamothe,
S. Tetrahedron Lett. 2001, 42, 2645-2648. (d) Fu¨rstner, A.; Dierkes, T.;
Thiel, O. R.; Blanda, G. Chem. Eur. J. 2001, 7, 5286-5298. For approaches
toward the benzolactone core, see: (e) Fu¨rstner, A.; Thiel, O. R.; Blanda,
G. Org. Lett. 2000, 2, 3731-3734. (f) Feutrill, J. T.; Holloway, G. A.;
Hilli, F.; Hu¨gel, H. M.; Rizzacasa, M. A. Tetrahedron Lett. 2000, 41, 8569-
8572. (g) Georg, G. I.; Ahn, Y. M.; Blackman, B.; Farokhi, F.; Flaherty,
P. T.; Mossman, C. J.; Roy, S.; Yang, K. J. Chem. Soc., Chem. Commun.
2001, 255-256.
(11) For the synthesis of apicularen A and derivatives, see: (a) Bhattacharjee,
A.; De Brabander, J. K. Tetrahedron Lett. 2000, 41, 8069-8073. (b)
Bhattacharjee, A.; Seguil, O. R.; De Brabander, J. K. Tetrahedron Lett.
2001, 42, 1217-1220. For an approach toward the benzolactone core, see:
(c) Lewis, A.; Stefanuti, I.; Swain, S. A.; Smith, S. A.; Taylor, R. J. K.
Tetrahedron Lett. 2001, 42, 5549-5552.
(5) Kim, J. W.; Shin-ya, K.; Furihata, K.; Hayakawa, Y.; Seto, H. J. Org.
Chem. 1999, 64, 153-155.
(12) The revised structure is shown.
(6) Boyd, M. R.; Farina, C.; Belfiore, P.; Gagliardi, S.; Kim, J.-W.; Hayakawa,
Y.; Beutler, J. A.; McKee, T. C.; Bowman, B. J.; Bowman, E. J. J.
Pharmacol. Exp. Ther. 2001, 297, 114-120.
(13) For model studies utilizing a conceptually similar approach see: (a) Snider,
B. B.; Song, F. Org. Lett. 2000, 2, 407-408. (b) Kuramochi, K.; Watanabe,
H.; Kitahara, T. Synlett 2000, 397-399. (c) Stefanuti, I.; Smith, S. A.;
Taylor, R. J. K. Tetrahedron Lett. 2000, 41, 3735-3738.
(7) For an extensive review, see: Forgac, M. AdV. Mol. Cell Biol. 1998, 23B,
403-453.
(14) For alternative approaches see refs 10b,c and: (a) Shen, R.; Porco, J. A.,
Jr. Org. Lett. 2000, 2, 1333-1336. (b) Fu¨rstner, A.; Brehm, C.; Cancho-
Grande, Y. Org. Lett. 2001, 3, 3955-3957.
(8) For a therapeutic focus, see: Farina, C.; Gagliardi, S. Drug DiscoVery Today
1999, 4, 163-172.
9
3246 J. AM. CHEM. SOC. VOL. 124, NO. 13, 2002

Synthesis of Salicylihalamide A and Related Compounds
A R T I C L E S
Scheme 3. Initial Synthesis of a C9-C17 Fragmenta
Scheme 2. Retrosynthetic Analysis
A connectivity analysis of the benzolactone core of salicy-
lihalamide A points to two equally attractive routes for its
assembly (Scheme 2). The first one features an esterification/
intramolecular olefin metathesis sequence (RCM) to form the
C9-C10 bond15 and offers the advantage of operational
simplicity combined with functional group tolerance (A1 + B1
f AB; Path A).16 Fu¨rstner and co-workers demonstrated the
feasibility of this approach to form a 12-membered resorcylic
benzolactone related to lasiodiplodin while avoiding isomer-
ization of the endocyclic double bond to the styryl position.17
Despite the robustness of the RCM in carbon-carbon bond
formation, it can only be implemented successfully for the
synthesis of salicylihalamides if concomitant stereocontrol for
the desired E-isomer can be exerted. As detailed in full later in
this paper, we have identified a highly E-selective RCM avenue
to salicylihalamide A.18 At the outset of our work, cross-
coupling19 of a stereodefined E-alkenyl organometallic fragment
B2 with a benzyl halide A2 was envisioned as a more robust
alternative and would join the C8-C9 bond with control and
maintenance of olefin geometry (Path B).20 Importantly, both
strategies converge to a common alkyne precursor B, adding
flexibility to the synthesis.
a Reagents and conditions: (a) Et2BOTf, i-Pr2NEt, CH2Cl2, -5 °C; then
add 3 at -78 °C (99% of a separable 85:15 diastereomeric mixture); (b)
4-MeOBnOC(dNH)CCl3, TfOH, Et2O, room temperature (20-80%); (c)
DIBAL-H, CH2Cl2, -78 °C; then MeOH, -78 °C f room temperature;
(d) 2b, TiCl4, i-Pr2NEt, CH2Cl2, -78 °C; then add 6 (74%, two steps); (e)
TBSOTf, 2,6-lutidine, CH2Cl2, -10 °C f room temperature (90%); (f)
LiEt3BH, THF, -78 °C f -40 °C (84%); (g) TsCl, Et3N, DMAP, CH2Cl2
(91%); (h) LiEt3BH, THF, room temperature (86%).
yields ranging from 20 to 80%.25 Aldehyde 6, obtained via a
one-step reduction of protected aldol derivative 5 (DIBAL-H,
CH2Cl2, -78 °C; MeOH quench at -78 °C),26 was subjected
to a stereoselective syn-aldol reaction with the titanium enolate27
derived from (2R)-N-(4-pentynoyl)bornanesultam 2b, and pro-
vided only one diastereoisomer 7 as judged by 1H NMR analysis
of the crude reaction mixture. Our major motivation to single
out an aldol tactic to form the C12-C13 bond was the
opportunity to introduce a molecular handle at a site that would
otherwise not be accessible for derivatization. Such a molecular
handle could prove extremely useful for probing reagent
development and/or analogue synthesis. For the completion of
the natural product, however, an oxidation state adjustment was
required. To this end, the â-hydroxy amide 7 was protected as
the tert-butyldimethylsilyl (TBS) ether 8 followed by reduction
of the N-acyl sulfonamide 8 (LiEt3BH), tosylate formation, and
further reduction of 10 (LiEt3BH) to complete the synthesis of
fragment 11.
Results and Discussion
The synthesis of alkyne 11 (fragment B) starts from readily
available aldehyde 3 as described in Scheme 3.21 Aldol reaction
of aldehyde 322 with the borylenolate derived from (2R)-N-
acetylbornanesultam 2a gave a separable mixture of two aldol
products, 4 and its epimer, in an 85:15 ratio.23 Protection of
the â-hydroxy amide 4 as the p-methoxybenzyl (PMB) ether 5
exploited the Bundle trichloroacetimidate protocol,24 but proved
to be highly capricious in our hands, providing irreproducible
With alkyne 11 in hand, the stage was set to explore its
assembly with the aryl sector. In one of two approaches (path
B, Scheme 2), this would involve a hydrometalation of the triple
bond followed by cross-coupling with benzylic bromide 1328
(eq 1). In this context, zirconocene 12a (M ) ZrCp2Cl) was
considered the most appealing nucleophilic coupling partner.
Indeed, stereodefined 1-(E)-alkenyl zirconocenes are readily
accessible via a functional group tolerant hydrozirconation of
1-alkynes and engage in cross-coupling chemistry with a variety
(15) Salicylihalamide numbering1 will be used throughout this paper.
(16) For selected reviews on ring-closing metathesis, see: (a) Schuster, M.;
Blechert, S. Angew. Chem., Int. Ed. Engl. 1997, 36, 2037-2056. (b) Grubbs,
R. H.; Chang, S. Tetrahedron 1998, 54, 4413-4450. (c) Wright, D. L.
Curr. Org. Chem. 1999, 3, 211-240. (d) Fu¨rstner, A. Angew. Chem., Int.
Ed. 2000, 39, 3012-3043.
(17) A modest selectivity for the E-isomer was observed (ratio 2.3:1): Fu¨rstner,
A.; Seidel, G.; Kindler, N. Tetrahedron 1999, 55, 8215-8230.
(18) After our total synthesis, all successful syntheses of salicylihalamide so
far have relied on a RCM approach: see refs 10a-d.
(19) Diederich, F.; Stang, P. J.; Eds. Metal-Catalyzed, Cross-Coupling Reactions;
Wiley-VCH: Weinheim, 1998.
(20) For an application of this approach in a model system see ref 10f.
(21) Because of the erroneous assignment of absolute stereochemistry to (-)-
salicylihalamide A, our explorative studies and initial total synthesis engaged
materials that reflect an antipodal relationship to those that would be
required for the natural enantiomer.
(22) Prepared in two steps from 1,3-propanediol: Eppley, A. W.; Totah, N. I.
Tetrahedron 1997, 53, 16545-16552.
(23) Bond, S.; Perlmutter, P. J. Org. Chem. 1997, 62, 6397-6400.
(24) (a) Nakajima, N.; Horita, K.; Abe, R.; Yonemitsu, O. Tetrahedron Lett.
1988, 29, 4139-4142. (b) Iversen, T.; Bundle, D. R. J. Chem. Soc., Chem.
Commun. 1981, 1240-1241.
(25) We later changed our protecting group strategy. At this point, enough
material could be obtained to continue exploratory work.
(26) Oppolzer, W.; Darcel, C.; Rochet, P.; Rosset, S.; De Brabander, J. HelV.
Chim. Acta 1997, 80, 1319-1337.
(27) Evans, D. A.; Rieger, D. L.; Bilodeau, M. T.; Urp´ı, F. J. Am. Chem. Soc.
1991, 113, 1047-1049.
(28) Hamada, Y.; Hara, O.; Kawai, A.; Kohno, Y.; Shiori, T. Tetrahedron 1991,
47, 8635-8652.
9
J. AM. CHEM. SOC. VOL. 124, NO. 13, 2002 3247

A R T I C L E S
Wu et al.
Scheme 4
of benzyl halides.29 We found that benzylic bromide 13
efficiently underwent Pd[PPh3]4 mediated cross-coupling with
the alkenyl zirconocene derivative in situ prepared by hydro-
zirconation of 1-heptyne (ClHZrCp2,30 THF). Unfortunately, all
attempts to obtain zirconocene 12a via hydrozirconation of
alkyne 11 were unsuccessful.31 In contrast, alkyne 11 did
undergo smooth hydrostannylation (Bu3SnH, AIBN, PhMe f
12b, M ) SnBu3) as a prelude to an alternative Stille coupling32
with benzylic bromide 13. This time, however, the correspond-
ing vinylstannane 12b was extremely prone to protodestann-
ylation, yielding terminal alkene 12d (M ) H) upon workup.33
After a lot of experimentation, the desired coupling product 14
could be obtained via a hydroboration (Sia2BH, THF f 12c,
M ) BSia2)/Suzuki cross-coupling sequence in 15% yield (13,
CsCO3, Ph3As, catalyst PdCl2(dppf)2, H2O/DMF).34 Again, the
culprit seemed to be an inefficient hydrometalation as starting
alkyne 11 accounted for the remaining mass balance.35
Scheme 5. Optimized Synthesis of the C10-C17 Fragmenta
At this point, the validity of planning a flexible strategy
became apparent and it was time to explore the RCM path to
salicylihalamides (Path A, Scheme 2). Toward this end, the
p-methoxybenzyl protecting group of alkene 12d was oxida-
tively (DDQ) deprotected (Scheme 4). Unfortunately, attempts
to acylate the resulting alcohol 15 with an activated ester derived
from benzoic acid derivative 16a36 (f 17a) uniformly met with
failure.37 A reactivity umpolung provided by a Mitsunobu
esterification was essential and delivered bis-olefin 17b in over
90% yield.38 Gratifyingly, the subsequent RCM with Grubbs’
ruthenium carbene complex39 produced the benzolactones in
essentially quantitative yield. A 4:1 mixture of separable isomers
resulted with the Z-isomer 18-Z predominating. However, this
material was improperly configured at C15, and as such, this
a Reagents and conditions: (a) catalyst OsO4, NMO, acetone-H2O; (b)
NaIO4, CH2Cl2 (97%, two steps); (c) 23, TiCl4, i-Pr2NEt, CH2Cl2, -78 °C;
then add 22, -78 °C (92%); (d) MOMCl, NaI, i-Pr2NEt, DME, reflux
(91%); (e) LiEt3BH, THF, -78 °C f room temperature (91%); (f) TsCl,
DMAP, Et3N, CH2Cl2, room temperature (91%); (g) LiEt3BH, THF, room
temperature (92%); (h) TBAF, THF, room temperature (98%); (i) 16a or
16b, DEAD, PPh3, Et2O, reflux (88-92%).
ratio does not per se reflect E/Z-selectivity for an eventual RCM
of a substrate en route to salicylihalamides. To address this point,
C15-epimeric compounds had to be prepared.
(29) Lipshutz, B. H.; Bulow, G.; Lowe, R. F.; Stevens, K. L. Tetrahedron 1996,
52, 7265-7276.
(30) In situ prepared ClHZrCp2 was critical for obtaining reproducible results:
Lipshutz, B. H.; Keil, R.; Ellsworth, E. L. Tetrahedron Lett. 1990, 31,
7257-7260.
Our initial route for the synthesis of the C10-C17 fragment
was plagued by a less than optimal diastereoselection during
the installation of the C15 stereocenter and an irreproducible
protection of the corresponding aldol product as the PMB
derivative (2 f 5, Scheme 3). Now was the time to correct
these imperfections, and a fully optimized gram-scale synthesis
of the C10-C17 fragment was developed (Scheme 5).40 First,
an enantioselective allylation41 of aldehyde 19 dramatically
improved the stereochemical purity at C15, providing the known
homoallyl alcohol 20 in 96% yield and 91% ee.42 Silylation
(TBSCl, imidazole, DMF) followed by oxidative double bond-
cleavage afforded aldehyde 22 reproducibly in 91% yield (3
(31) As a control experiment, a solution of in situ prepared HClZrCp2 was
divided into two portions to which were added 11 and 1-heptyne,
respectively. 1-Heptyne underwent smooth hydrozirconation as judged by
the subsequent cross-coupling with benzyl bromide 13. Alkyne 11 on the
contrary was recovered unchanged.
(32) For reviews see: (a) ref 19, pp 167-202. (b) Farina, V.; Krishnamurthy,
V.; Scott, W. J. Org. React. 1997, 50, 1-652.
(33) Cross-coupling with the crude vinylstannane mixture was ineffective.
(34) (a) Miyaura, N.; Yano, T.; Suzuki, A. Tetrahedron Lett. 1980, 21, 2865-
2868. (b) Miyaura, N.; Ishiyama, T.; Sasaki, H.; Ishikawa, M.; Satoh, M.;
Suzuki, A. J. Am. Chem. Soc. 1989, 111, 314-321.
(35) A variety of hydroborating agents were tested, including 9-BBN, cat-
echolborane, pinacolborane, freshly prepared Cy2BH, and freshly prepared
disiamylborane. At present, we can offer no explanation as to why this
particular alkyne resists hydrozirconation or hydroboration.
(36) For the preparation of 16a and 16b, see the Supporting Information.
(37) The following reaction conditions were unsuccessful: (1) in situ activation
of acid with TBTU, EDC, or DCC in combination with a variety of catalysts
(DMAP, HOBt); (2) in situ activation of acid with 2,4,6-trichlorobenzoyl
chloride; (3) direct acylation with the corresponding benzoyl chloride.
(38) Mitsunobu, O. Synthesis 1981, 1-28.
(40) From this point forward (except Figure 2), the stereochemistry drawn
corresponds to that of natural (-)-salicylihalamide A.
(41) Brown, H. C.; Randad, R. S.; Bhat, K. S.; Zaidlewicz, M.; Racherla, U. S.
J. Am. Chem. Soc. 1990, 112, 2389-2392.
(39) Schwab, P.; Grubbs, R. H.; Ziller, J. W. J. Am. Chem. Soc. 1996, 118,
100-110.
(42) Smith, A. B., III; Minbiole, K. P.; Verhoest, P. R.; Schelhaas, M. J. Am.
Chem. Soc. 2001, 123, 10942-10953.
9
3248 J. AM. CHEM. SOC. VOL. 124, NO. 13, 2002

Synthesis of Salicylihalamide A and Related Compounds
A R T I C L E S
Table 1. Ring-Closing Olefin Metathesis Studies (Eq 2)
transformation, albeit with different degrees of E/Z-selectivity.
Also, both differentially protected substrates 28a,b gave identical
E/Z-ratios under identical reaction conditions.47 Whereas pre-
catalyst 31 provided a 9:1 (E:Z) ratio with only a slight erosion
of selectivity over time, precatalyst 32 produced a lower 67:33
ratio, which remained constant over time.
products (ratio)a,b
substrate
t (h)
catalyst 31
catalyst. 32
28a
28a
28a
28b
28b
28b
1.3
6.3
18.3
1.3
6.3
18.3
29a:30a (90:10)
29a:30a (88:12)
29a:30a (80:20)
29b:30b (90:10)
29b:30b (88:12)
29b:30b (84:16)
29a:30a (66:34)
29a:30a (68:32)
29a:30a (67:33)
29b:30b (64:36)
29b:30b (67:33)
29b:30b (68:32)
In principle, olefin metathesis will produce a thermodynamic
distribution of products if secondary metathetical isomerizations
compete on the time scale of the experiment.48 There are two
primary factors that will affect the efficiency of secondary
metathetical isomerizations: (1) the activity of the propagating
Ru-methylidene species and (2) catalyst decomposition rates.49,50
In light of this, the more stable and more active “second
generation” catalysts (e.g. 32) were shown to enrich initially
formed products to the thermodynamic equilibrium ring closure
product.48d-g In contrast, kinetic product ratios cannot be ruled
out with the “first generation” Ru-alkylidene catalysts due to
their shorter lifetime (thermal instability)50 and less efficient
reaction with 1,2-disubstituted olefins (reaction products).16 We
conclude from our results that RCM with catalyst 32 rapidly
produces a thermodynamic ratio of products 29-30 on a time
scale that consumed all the starting material based on the
following observations: (1) the product ratio did not change
after prolonged exposure indicating that equilibrium was
established, (2) an identical product ratio is observed for the
formation of benzolactones 29-30 from both our precursors
(28a,b) and the nonidentical precursors described by Fu¨rstner
et al.,51 and (3) upon exposure of either geometrically pure 29a
or 30a to catalyst 32, an identical 67:33 mixture of 29a:30a
was formed. In contrast, the first generation Grubbs’ catalyst
31 kinetically induced the formation of the desired E-isomers
29a,b with relatively high selectivity and a thermodynamic
product ratio was never reached, even after prolonged reaction
times.
a Ratio determined by 1H NMR analysis. b In each individual experiment,
the combined isolated yield was >93%.
steps). Treatment of this aldehyde with the Z(O)-titanium enolate
derived from (2S)-N-(4-pentenoyl) bornanesultam 2343 produced
exclusively one diastereomeric aldol product 24 in 92% yield.
As before (Scheme 3), a three-step sequence starting from the
methoxymethyl ether 25 accomplished the oxidation state
readjustment to deliver the C12 methyl-substituted fragment 26.
Fluoride-assisted liberation of the C15 alcohol and subsequent
Mitsunobu38 coupling of 27 with carboxylic acid 16a36 or the
corresponding phenolic MOM ether 16b36 set the stage for a
detailed study of the crucial ring-closing olefin metathesis of
bis-olefins 28a,b.
The lack of a stereopredictive model for the formation of
large rings via RCM is exemplified by our results with the
metathetical ring closure of substrates 28a,b. Whereas the
diastereomeric substrate 17b gave the Z-olefin 18-Z (vide supra,
Scheme 4) as the major isomer, 28a fortuitously produced the
E-benzolactone 29a with an impressive selectivity of 9:1 when
subjected to similar reaction conditions (eq 2; 5 mol % catalyst
Having secured a viable sequence to the benzolactone core
of salicylihalamide A, we turned our attention to the installation
of the acylated enamine side chain. Toward this end, the
p-methoxybenzyl ether in 29a was oxidatively removed (DDQ)
and the resulting alcohol 33 was oxidized with Dess-Martin
periodinane (Scheme 6).52 Engagement of the resulting aldehyde
34 in a Horner-Wadsworth-Emmons (HWE) homologation
with trimethyl phosphonoacetate provided methyl ester 36a as
an inseparable mixture of E/Z-isomers in a ratio of 4:1. After
31, CH2Cl2, room temperature).44 To confuse the issue even
more, an initial single experiment with the corresponding
phenolic methoxymethyl ether 28b furnished benzolactones 29b
and 30b with an eroded selectivity of 3:1 upon exposure to
catalyst 31. In light of the above, a detailed study of the RCM
of 28a,b with Ru-alkylidene precatalysts 31 and 3245 was
conducted (eq 2 and Table 1).46 It is clear from these results
that both catalysts are equally efficient in performing the desired
(47) This indicates that the differences in E/Z-selectivity for the RCM of 28a
vs 28b observed during our initial experiments with catalyst 31 must be
due to subtle variations in reaction conditions.
(48) For selected examples of reversible olefin metathesis reactions, see: (a)
Marsella, M. J.; Maynard, H. D.; Grubbs, R. H. Angew. Chem., Int. Ed.
Engl. 1997, 36, 1101-1103. (b) Xu, Z.; Johannes, C. W.; Houri, A. F.;
La, D. S.; Cogan, D. A.; Hofilena, G. E.; Hoveyda, A. H. J. Am. Chem.
Soc. 1997, 119, 10302-10316. (c) Hamilton, D. G.; Feeder, N.; Teat, S.
J.; Sanders, J. K. M. New J. Chem. 1998, 22, 1019-1021. (d) Smith, A.
B., III; Adams, C. M.; Kozmin, S. A.; Paone, D. V. J. Am. Chem. Soc.
2001, 123, 5925-5937. (e) Lee, W. C.; Grubbs, R. H. Org. Lett. 2000, 2,
2145-2147; correction: Org. Lett. 2000, 2, 2559. (f) Fu¨rstner, A.; Thiel,
O. R.; Ackermann Org. Lett. 2001, 3, 449-451. (g) Lee, W. C.; Grubbs,
R. H. J. Org. Chem. 2001, 66, 7155-7158. (h) Wright, D. L.; Usher, L.
C.; Estrella-Jimenez, M. Org. Lett. In press.
(49) For the mechanism and activity of ruthenium olefin metathesis catalysts,
see: Sanford, M. S.; Love, J. A.; Grubbs, R. H. J. Am. Chem. Soc. 2001,
123, 6543-6554 and references therein.
(50) For stability and decomposition studies, see: Ulman, M.; Grubbs, R. H. J.
Org. Chem. 1999, 64, 7202-7207.
(43) Oppolzer, W.; Osamu, T.; Deerberg, J. HelV. Chim. Acta 1992, 75, 1965-
1978.
(44) For selected examples of remote functionality affecting stereoselectivity,
see: (a) Fu¨rstner, A.; Langemann, K. J. Org. Chem. 1996, 61, 3942-
3943. (b) Meng, D.; Su, D.; Balog, A.; Bertinato, P.; Sorensen, E. J.;
Danishefsky, S. J.; Zheng, Y.; Chou, T.; He, L.; Horwitz, S. B. J. Am.
Chem. Soc. 1997, 119, 2733-2734.
(45) N-Heterocyclic carbene ligated ruthenium alkylidene catalysts have been
independently reported by three groups, cf.: (a) Huang, J.; Stevens, E. D.;
Nolan, S. P.; Petersen, J. L. J. Am. Chem. Soc. 1999, 121, 2674-2678. (b)
Scholl, M.; Trnka, T. M.; Morgan, J. P.; Grubbs, R. H. Tetrahedron Lett.
1999, 40, 2247-2250. (c) Ackermann, L.; Fu¨rstner, A.; Weskamp, T.; Kohl,
F. J.; Herrmann, W. A. Tetrahedron Lett. 1999, 40, 4787-4790.
(46) Identical reaction conditions (solvent, concentration, catalyst loading,
temperature, same batch of catalyst) were employed for all reactions of
Table 1.
(51) Fu¨rstner and co-workers recently reported results concerning the RCM of
compounds that differ from 28a,b only by an additional gem-dimethyl
substitution at the C9 olefin terminus.10d,e Limited to the use of a more
competent second-generation catalyst, compounds 29a/30a and 29b/30b
were obtained in an identical ratio of 67:33.
(52) Dess, D. B.; Martin, J. C. J. Am. Chem. Soc. 1991, 113, 7277-7287.
9
J. AM. CHEM. SOC. VOL. 124, NO. 13, 2002 3249

A R T I C L E S
Wu et al.
Scheme 6 a
Scheme 7. Completion of the Synthesis
a Reagents and conditions: (a) DDQ, CH2Cl2-H2O, room temperature
(96%); (b) Dess-Martin periodinane, CH2Cl2, room temperature (95%);
(c) 35a-c, NaH, THF, 0 °C (85-95%); (d) BBr3, CH2Cl2, -78 °C (81%);
(e) TBSCl, imidazole, DMAP, DMF, room temperature (92% of 38); (f)
Ba(OH)2‚8H2O, THF, room temperature; (g) TBSCl, imidazole DMAP,
DMF, room tmeperature (50% of 39, two steps); (h) catalyst Pd(PPh3)4,
morpholine, THF, room temperature (96%); (i) (PhO)2P(O)N3, Et3N, PhH,
room temperature (92-96%).
removal of the ether protecting groups with BBr3, the corre-
sponding E-methyl ester 37a was separated by flash chroma-
tography,53 hydrolyzed (Ba(OH)2‚8H2O) and extensively sily-
lated with excess TBSCl to deliver carboxylic acid 39.54 The
overall yield for the transformation of 34 to 39 was a
disappointing 20-30%. Therefore, a series of optimization
experiments were performed that quickly led to the use of allyl
diethylphosphonoacetate, delivering allyl ester 36b with 18:1
selectivity for the E-isomer.55 Subsequent deprotection, bis-
silylation, and Pd-catalyzed removal of the allyl ester improved
the overall yield of 39 dramatically (67% from aldehyde 34).
Transformation into acyl azide 40 ((PhO)2P(O)N3, NEt3, PhH)56
preceded a Curtius rearrangement induced by heat (PhH, 80
°C, 6 h). Although the corresponding isocyanate 41 was stable
to chromatographic purification, it was usually engaged in the
next reaction without purification (Scheme 7).
Final carbon-carbon bond formation proceeded smoothly via
the addition of hexadienyllithium, prepared in situ from bromide
4257 via metal-halogen exchange (t-BuLi, Et2O), to a -78 °C
solution of isocyanate 41 (Scheme 7). The desired compound
was obtained as an inseparable mixture of 22Z and 22E isomers
together with a chromatographically (silicagel) more mobile
fraction. Careful fluoride-assisted desilylation with a buffered
solution of commercially available HF‚pyridine in THF/pyridine
provided geometrical isomers 1a/1c, which could be separated
Figure 2. Rasmol Representation of the X-ray Structure of p-Bromoben-
zoate 47.
by semipreparative HPLC.58 The chromatographically more
mobile fraction was treated in a similar fashion and provided a
mixture of the corresponding salicylihalamide dimers 43/44,
again separable by semipreparative HPLC.59
In contrast to the current study, we had initially synthesized
ent-1a based on the absolute configuration reported for the
natural product.9a Although this synthetic material was found
to be identical to natural salicylihalamide A according to NMR
([D6]benzene and [D4]methanol), IR, UV, and coelution on
HPLC (two different solvent systems), and TLC (three different
solvent systems),60,61 the signs of the optical rotations of
synthetic ent-1a ([R]23D +20.8; c 0.12; MeOH) and the natural
product (reported: [R]23 -35; c 0.7; MeOH) were opposite.
D
Moreover, synthetic ent-1a was devoid of growth inhibitory
activity when screened against the NCI 60-cell line panel.61 At
this point we were fortunate that p-bromobenzoate derivative
47 provided crystals suitable for X-ray diffraction studies,
confirming the absolute configuration of our synthetic lactones
(Figure 2).9a Based on all the available evidence, the absolute
configuration of natural (-)-salicylihalamide A was assigned
(58) The ratios of 1a:1c and 43:44 are indicative of the isomeric ratio of bromide
42.
(53) The minor Z-isomer would be the desired one for the synthesis of
salicylihalamide B.
(54) The phenolic methyl ether and MOM ether had to be replaced with
protecting groups that could be removed after installation of the acid-labile
acylated enamine functionality.
(55) The use of tert-butyl dimethylphosphonoacetate 35c provided tert-butyl
ester 36c with 20:1 E:Z-selectivity. However, 36c decomposed to uniden-
tifiable products upon BBr3 treatment (CH2Cl2, -78 °C).
(56) Ninomiya, K.; Shioiri, T.; Yamada, S. Tetrahedron 1974, 30, 2151-2157.
(57) (1Z,3Z)-1-Bromo-1,3-hexadiene (42) was contaminated with 20-50% of
(1Z,3E)-1-bromo-1,3-hexadiene and used as such. For the preparation of
42, see the Supporting Information.
(59) These dimers are presumably formed through reaction of the intermediate
lithiated amide with a second molecule of isocyanate 41. For a similar
observation, see ref 13b.
(60) Careful examination of the 1H NMR spectra of natural salicylihalamide A
indicates the presence of a minor contaminant with an identical spectro-
scopic signature to our synthetic Z,E-isomer 1c (see spectra provided in
the Supporting Information). This contaminant could be a natural isomer
of salicylihalamide A, or the result of isomerization during the isolation/
purification procedure.
(61) We thank Dr. Michael Boyd (NCI) for providing a sample of natural
salicylihalamide A and for testing synthetic 1a,c in the 60-cell line screen.
9
3250 J. AM. CHEM. SOC. VOL. 124, NO. 13, 2002

Synthesis of Salicylihalamide A and Related Compounds
A R T I C L E S
Scheme 8. Synthesis of Side Chain Modified Salicylihalamidesa
A minimally perturbed side chain that would retain the
potentially hydrogen bonding characteristics (donor and/or
acceptor) of the N-acyl functionality was envisioned to arise
from a simple saturation of the enamine double bond of
biologically active (vide infra) salicylihalamide derivative 50.
However, direct hydrogenation of 50 also saturated the endo-
cyclic double bond to produce 59 (Scheme 9). Because there
was no obvious short solution to this chemoselectivity problem,
a control reagent 58 was prepared as a probe to investigate
independently the effect of endocyclic double bond saturation
on biological activity (Scheme 9). Our point of departure
entailed a hydrogenation of 29a or 30a with concomitant
removal of the p-methoxybenzyl (PMB) ether. Subsequent
conversion of 54 to 58 took full advantage of the chemistry
outlined for the preparation of 50 without complication.
Hydrogenation of this material also yielded the fully saturated
salicylihalamide derivative 59.
Although the potent in vivo and in vitro biological activity
of salicylihalamide-based dimers 43-45 (vide infra) pointed
to a potential site for attachment of a photoactivatable radio-
active probe, the phenolic and secondary hydroxyls were also
investigated as a handle for derivatization (Scheme 10). Starting
with bis-TBS derivative 38, selective deprotection of the
phenolic (TBAF, THF, 0 °C, 91%) or secondary TBS ether
(aqueous HCl, 91%) was followed by benzoylation to furnish
benzoates 60 (78%) and 61 (50%), respectively. Together with
compound 36b, these materials were independently elaborated
to bis- and monoprotected forms (62-64) of salicylihalamide
derivative 50.
The in vitro inhibition of V-ATPase activity by salicylihal-
amide A, its enantiomer, and the other synthetic derivatives is
summarized in Table 2. For our studies, we utilized a recon-
stituted, fully purified V-ATPase from bovine brain62 to
completely eliminate potential effects arising from contaminating
nonvacuolar ATPase activities present in the crude membrane
preparations utilized in the initial6 study. Here, we demonstrate
for the first time that synthetic (-)-salicylihalamide A inhibits
ATP-energized proton pumping of the intact, reconstitutively
active V-ATPase with an IC50 value of <1.0 nM, as compared
to 3.1 nM for bafilomycin A (Table 2, entries 1-2).63 The
unnatural enantiomer (+)-salicylihalamide A was 300-fold less
potent (Table 2, entry 3), further confirming the absolute
configuration of natural salicylihalamide A.
a Reagents and conditions: (a) phenylacetylene or 1-hexyne, n-BuLi,
-78 °C; then add 41, -78 °C; (b) HF‚Py, Py-THF, room temperature (48:
57%, 49: 50%; two steps); (c) 1-pentanol or 1-pentanethiol or farnesol,
40, PhH, 80 °C; (d) HF‚Py, Py-THF, room temperature (50: 47%, 51: 80%,
52: 50%; two steps).
to be as drawn in 1a. Unequivocal proof for the absolute
stereochemistry of natural salicylihalamide ultimately came from
biological characterization of synthetic 1a, which provided a
differential cytotoxicity profile undistinguishable from the
natural product in the NCI-60 cell line panel.61
With issues related to absolute stereochemistry resolved, we
decided to initiate a detailed study to unravel salicylihalamide’s
mode-of-action at the cellular and biochemical level. During
the course of this work, the Vacuolar ATPase (V-ATPase) was
identified as a putative target of salicylihalamide A.6 In contrast
to bafilomycins and other prototypical V-ATPase inhibitors,
salicylihalamide A exquisitely discriminates between mam-
malian and nonmammalian V-ATPases.6 To fully exploit this
unique feature of salicylihalamide, it is imperative to map its
binding site on V-ATPase in molecular detail. To identify
regions in the molecule that would tolerate the obligate structural
changes that will accompany the introduction of the prerequisite
probes, we prepared a variety of salicylihalamide analogues.
Ideally, such reporter constructs would be available from late
stage intermediates, avoiding extensive chemical sequences. Our
initial efforts were therefore orchestrated toward side chain
modifications, emanating from a common, naturally configured
isocyanate 41, now accessible in 20 steps (longest linear
sequence) and 20% overall yield. Thus, compound 46 and the
corresponding dimer 45 were prepared by addition of hexyl-
lithium (instead of hexadienyllithium) to isocyanate 41 followed
by final deprotection (Scheme 7). An inverse addition of
isocyanate 41 to a cold solution (-78 °C) of organolithium
nucleophiles was next explored to suppress dimer formation.59
Indeed, preparation of alkynoyl enamine derivatives 48/49
followed this procedure, and no trace of dimer formation was
detected (Scheme 8). Alternatively, carbamates or thiocarbam-
ates 50-52 are obtained in an operationally simplified procedure
by heating the acyl azide 40 in the presence of the appropriate
alcohol or thiol. In light of the fact that V-ATPase is a
membrane-bound protein, we also prepared analogue 52 incor-
porating a lipophilic farnesyl anchor hoping to target this
derivative to the membrane (Scheme 8).
Side chain modified analogues that lack salicylihalamide’s
characteristic N-acyl enamine functionality are attractive can-
didates for the following reasons: (1) they are expected to confer
increased acid stability; (2) they can potentially be prepared
via shorter sequences; and (3) they would answer an important
question related to the functional role of the N-acyl enamine
moiety. Octanoate 53, a compound with identical chain length
and hydrophobicity similar to those of salicylihalamide, is
representative of this class of compounds and was prepared from
alcohol 33 via a Mitsunobu38 esterification followed by depro-
tection (eq 3).
Side chain modified analogues 43-46 and 48-51 all retain
the ability to potently inhibit proton pumping activity at
(62) (a) Xie, X.-S.; Tsai, S.-J.; Stone, D. K. Proc. Natl. Acad. Sci. U.S.A. 1986,
83, 8913-8917. (b) Mattsson, J. P.; Schlesinger, P. H.; Keeling, D. J.;
Teitelbaum, S. L.; Stone, D. K.; Xie, X.-S. J. Biol. Chem. 1994, 269,
24979-24982.
(63) In contrast to the initial study,6 which measures inhibition of ATPase
activity, we utilized a H+ transport (pumping) assay that is a far more
specific and accurate measure for V-ATPases than the ATPase assay. The
pumping assay requires every part of the enzyme functioning. The ATPase
activity, on the other hand, only tells its partial activity, i.e., ATP hydrolysis,
which may or may not be coupled to proton pumping. For a description of
the proton pumping assay, see: Crider, B. P.; Xie, X.-S.; Stone, D. K. J.
Biol. Chem. 1994, 269, 17379-17381.
9
J. AM. CHEM. SOC. VOL. 124, NO. 13, 2002 3251

A R T I C L E S
Wu et al.
Scheme 9. Synthesis of Reduced Salicylihalamidesa
modated without abrogating biological activity (43-45). How-
ever, there is a size restriction associated with the N-acyl
fragment indicated by the 1000-fold drop in activity of farne-
sylated analogue 52. This result also demonstrates that the side
chain is not merely functioning as a lipophilic membrane anchor.
The seriously compromised potency of octanoate 53 and allyl
ester 37b in the in vitro V-ATPase assay demonstrated the
importance of the N-acyl enamine functionality.64 Moreover,
whereas the N-carbamoyl enamine derivative 58 retained a
significant capacity to inhibit proton pumping in the V-ATPase
assay, the enamide to amine permutation (58f59) substantially
attenuated inhibitory potential. Also, benzoylation of the C15
alcohol (64) or the C3 phenol (63) led to a significant drop in
biological activity, thereby virtually eliminating the possibility
of exploiting the most obvious functional handles to develop a
useful salicylihalamide-based probe reagent.
Initially, we guided our chemical work with a cell-based assay
looking for growth inhibition of the SK-MEL 5 human
melanoma cell line.65 Although the structure-activity relation-
ships found in this assay [IC50 (µM): 1a/c, 0.06; 43, 0.04; 44,
0.1; 45, 0.6, 46, 0.38; 48, 0.3; 49, 0.3; 50, 0.5; 51; 0.45; 52,
1.5; 53, >20; 37b, >20; 58, 8; 59, >20, 63, 1, 64, >20] mirror
those of the in vitro V-ATPase assay, they do not per se indicate
a link between inhibition of the vacuolar ATPase and cytotox-
icity.66 Further biochemical and cell biological studies will be
necessary to implicate modulation of V-ATPase function or the
function of an alternative cellular target as the mechanism by
which salicylihalamides induce differential cytotoxicity in
cultured mammalian cells. These studies will require the
development of relevant fluorescent and/or radioactive photo-
activatable probe reagents. The structure-function data pre-
sented here point to the side chain nitrogen as a point for
attachment or modification of the N-acyl terminus as the most
promising avenue for obtaining these probe reagents.
a Reagents and conditions: (a) 29a, 10% Pd/C, H2, MeOH, room
temperature (82%); or 30a, 10% Pd/C, NH4O2CH, MeOH, reflux (67%);
(b) Dess-Martin periodinane, CH2Cl2, room temperature (81%); (c)
CH2dCHCH2OP(O)(OEt)2, NaH, THF, 0 °C (93%); (d) BBr3, CH2Cl2, -78
°C (97%); (e) TBSCl, imidazole, DMAP, DMF, room temperature (67%);
(f) catalyst Pd(PPh3)4, morpholine, THF, room temperature (92%); (g)
(PhO)2P(O)N3, Et3N, PhH, room temperature; (h) 1-pentanol, PhH, 80 °C;
(i) HF‚Py, Py-THF, room temperature (43%; three steps); (j) 10% Pd/C,
H2, MeOH, room temperature (25%).
Scheme 10. Synthesis of Hydroxy-Derivatized Salicylihalamidesa
Conclusions
Herein, we have detailed in full our synthetic efforts toward
salicylihalamide A, the first example of a structurally novel class
of natural products that induce unique biological effects in
cultured mammalian cells. The synthesis of the structure
originally assigned to natural salicylihalamide A as well as the
corresponding enantiomer have unambiguously established the
absolute configuration of the natural product as presented by
1a. The advantage of a flexible synthetic strategy was high-
lighted by our ability to quickly advance the ring-closing olefin
metathesis approach as the method of choice to build the 12-
membered benzolactone core of salicylihalamides. Notably, a
detailed study of this key step identified conditions that
kinetically favor formation of the desired E-benzolactone 29a
with high levels of stereocontrol. The end game included
elaboration of this material to a stereodefined E-alkenyl iso-
cyanate 41 through Horner-Wadsworth-Emmons homologa-
tion of aldehyde 34, followed by Curtius rearrangement of acyl
azide 40. Final addition of hexadienyllithium to a cold solution
of vinyl isocyanate 41, followed by a mild deprotection delivered
a Reagents and conditions: (a) TBAF, THF, room temperature (84%);
(b) BzCl, Py, room temperature (78%); (c) concentrated HCl, THF, room
temperature (91%); (d) BzCl, Py, room temperature (50%); (e) catalyst
Pd(PPh3)4, morpholine, THF, room temperature (85-97%); (f) (PhO)2P(O)N3,
Et3N, PhH, room temperature (68-96%); (h) 1-pentanol, PhH, 80 °C (62:
55%; benzoates: not purified) (i) HF‚Py, Py-THF, room temperature (63:
55%, 64: 33%; two steps).
Table 2. Inhibition of the H+ Transport Activity of Reconstituted
V-ATPase from Bovine Brain by Synthetic Salicylihalamidesa
compd
IC50 (nM)
compound
IC50 (nM)
Bafilomycin A
1a
ent-1a
37b
43
44
45
46
48
3.1
<1
270
230
3.7
1.2
3
49
50
51
52
53
58
59
63
64
1
1.6
1.8
>1000
1000
3
30
300
180
1
<1
a IC50 values (measured according to ref 63) are the average of at least
three experiments (error (5%).
(64) However, an intact side chain is not sufficient for biological activity given
the inactivity of the unnatural enantiomer of salicylihalamide A.
(65) This cell line was reported to be particularly sensitive to natural salicyli-
halamide A.1
(66) A detailed study of the cytotoxic effects of salicylihalamide A and analogues
against a panel of human breast and lung cancer cell lines, including drug-
resistant cell lines, will be published elsewhere.
concentrations similar to those of the parent compound. This
indicates that the hexadienoyl fragment is not crucial for
biological activity and that substantial sterically demanding
modifications emanating off the amide nitrogen can be accom-
9
3252 J. AM. CHEM. SOC. VOL. 124, NO. 13, 2002

Synthesis of Salicylihalamide A and Related Compounds
A R T I C L E S
salicylihalamide A (1a), the 22E-isomer 1c and the correspond-
ing salicylihalamide dimers 43 and 44.
grants CA 90349 (J. K. De Brabander) and DK 33627 (X.-S.
Xie), and junior faculty awards administered through the
Howard Hughes Medical Institute and the University of Texas
Southwestern Medical Center are gratefully acknowledged.
HRMS analyses were performed at the NIH regional mass
spectrometry facility at the University of Washington,; St. Louis,
MO. J. K. De Brabander is a fellow of the Alfred P. Sloan
Foundation.
The highly efficient synthesis of acyl azide 40 (20-25%
overall yield; 19 steps longest linear sequence from aldehyde
19) was a crucial asset in investigating salicylihalamide
structure-activity relationships. A collection of 17 salicylihal-
amide-based structures have been synthesized and evaluated as
inhibitors of the Vacuolar (H+)-ATPase. These investigations
have identified several highly potent analogues of salicylihal-
amide A, and provided the framework for the development of
probe reagents. These reagents will be valuable tools for
mapping the salicylihalamide-binding site on V-ATPase and
investigating the mechanism(s) by which salicylihalamide
induces differential cytotoxicity in mammalian cells.
Supporting Information Available: Experimental procedures
and characterization data and NMR spectra of synthetic and
natural salicylihalamide and compounds 43-46, 48-52, 58-
59, and 62-64 (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
Acknowledgment. Financial support provided by the Robert
A. Welch Foundation, the National Institutes of Health through
JA0177713
9
J. AM. CHEM. SOC. VOL. 124, NO. 13, 2002 3253