Total Synthesis and Biological Evaluation of Halipeptins A
and D and Analogues
K. C. Nicolaou,* Dimitrios E. Lizos, David W. Kim, Daniel Schlawe,
Rita G. de Noronha, Deborah A. Longbottom, Manuela Rodriquez,†
Mariarosaria Bucci,‡and Giuseppe Cirino‡
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 January 4, 2006; E-mail: email@example.com
Abstract: The marine-derived halipeptins A (1a) and D (1d) and their analogues 3a, 3d and 4a, 4d were
synthesized starting from building blocks 10, 13, 14a or 14d, 15, and 16. The first strategy for assembling
the building blocks, involving a macrolactamization reaction to form the 16-membered ring hydroxy thioamide
52d as a precursor, furnished the epi-isoleucine analogue (4d) of halipeptin D, whereas a second approach
involving thiazoline formation prior to macrolactamization led to a mixture of halipeptins A (1a) and D (1d)
and their analogues 3a, 3d (epimers at the indicated site) and 4a, 4d (epimers at the indicated site). The
same route starting with D-Ala resulted in the exclusive formation of the epimeric halipeptin D analogue
3d. The synthesized halipeptins, together with the previously constructed oxazoline analogues 5d and 6d,
were subjected to biological evaluation revealing anti-inflammatory properties for 1a, 1d, and 6d while
being noncytotoxic against human colon cancer cells (HCT-116).
The halipeptins (A-D, Figure 1) are a group of marine-
derived depsipeptides with interesting molecular architectures
and biological properties. Their fascinating tale began with the
isolation of halipeptins A and B by the Gomez-Paloma group
in 2001 from the sponge Haliclona sp., and the assignment to
them of depsipeptide structures containing an intriguing 1,2-
oxazetidine-type structural motif as shown in 2a and 2b (Figure
1).1Adding to the attractiveness of these molecules as targets
for total synthesis was the potent anti-inflammatory activity
attributed to the most abundant of the two, halipeptin A (2a).
The latter compound exhibited 60% reduction of carrageenan-
induced paw edema in mice at the intraperitoneal dose of 0.3
mg kg-1body weight, thus rivaling in potency commercial anti-
The story of halipeptins took another turn in 2002 when the
same group reported a third member of the family, halipeptin
C (1c, Figure 1), and revised their originally proposed oxaze-
tidine structures of halipeptins A and B to the thiazoline
structures 1a and 1b, respectively (Figure 1).2Adding support
to the new structures was yet another development in the field,
that involving the isolation of halipeptin D (1d, Figure 1) from
a different sponge, Leiosella cf. arenifibrosa, by Faulkner and
Manam.3In 2003, these researchers communicated to us the
structure of halipeptin D (1d) and their finding that their newly
isolated natural product exhibited potent cytotoxic properties
[IC50) 7 nM against human colon cancer HCT-116 cell line
and an average IC50) 420 nM against a BMS ODCP (oncology
diverse cell panel) of tumor cell lines].4These results were in
sharp contrast to those of Gomez-Paloma for their halipeptins
A-C (1a-1c), which apparently were devoid of any significant
cytotoxicity properties.1,2It should also be mentioned that neither
the Gomez-Paloma group nor the Faulkner-Manam team could
assign with certainty the absolute stereochemistries at C-3 and
C-4, although the former group had reached the conclusion,
through chemical synthesis studies, that the absolute stereo-
chemistry of the C-7 stereocenter was of the (S) configuration
in halipeptin B (1b).1
Faced with these intriguing and still unresolved questions and
because of the natural scarcity of these compounds, we
embarked on a program directed toward their total synthesis,
initially targeting halipeptins A (1a) and D (1d).5In this Article,
we describe in detail our investigations in this area, including
the chemical synthesis of these two natural products (1a and
†Current address: Dipartimento di Scienze Farmaceutiche, University
degli Studi di Salerno, Italy.
‡Current address: Dipartimento di Farmacologia Sperimentale, Univer-
sity degli Studi di Napoli, Italy.
(1) Randazzo, R.; Bifulco, G.; Giannini, C.; Bucci, M.; Debitus, C.; Cirino,
G.; Gomez-Paloma, L. J. Am. Chem. Soc. 2001, 123, 10870-10876.
(2) Della Monica, C.; Randazzo, A.; Bifulco, G.; Cimino, P.; Aquino, M.; Izzo,
I.; De Riccardis, F.; Gomez-Paloma, L. Tetrahedron Lett. 2002, 43, 5707-
(3) Nicolaou, K. C.; Schlawe, D.; Kim, D.; Longbottom, D. A.; de Noronha,
R. G.; Lizos, D. E.; Rao Manam, R.; Faulkner, D. J. Chem.-Eur. J. 2005,
(4) (a) Izzo, I.; Avallone, E.; Della Corte, L.; Maulucci, N.; De Riccardis, F.
Tetrahedron: Asymmetry 2004, 15, 1181-1186 (for some biological data
of halipeptin D, see ref 3 in this publication). (b) Data obtained by D. J.
Faulkner and which appeared in a summary review, see: Fenical, W.; et
al. Pharm. Biol. 2003, 41, 6-14.
(5) Nicolaou, K. C.; Kim, D.; Schlawe, D.; Lizos, D. E.; de Noronha, R. G.;
Longbottom, D. Angew. Chem., Int. Ed. 2005, 44, 4925-4929.
Published on Web 03/15/2006
4460 9 J. AM. CHEM. SOC. 2006, 128, 4460-4470
10.1021/ja060064v CCC: $33.50 © 2006 American Chemical Society
1d), their epimers 3a, 3d, 4a, and 4d (Figure 1), and their
biological properties, as well as those of a number of oxazoline
analogues of halipeptin D (5d3and 6d3Figure 1).
Results and Discussion
In contemplating a total synthesis of the halipeptins, a number
of their structural features deserve special attention and dictate
certain strategies and tactics. A detailed discussion of their
structural motifs is, therefore, deemed important at this juncture.
Thus, even a cursory inspection of the structures of halipeptins
reveals a striking number of methyl groups situated on, or
around, the periphery of their macrocyclic depsipeptide ring.
These methyl groups translate into an unusually high degree of
strain for a 16-membered ring that otherwise could have been
relatively free of such energetic barriers, as far as its construction
is concerned. Further constrains are imposed upon this macro-
cycle by the thiazoline ring and its adjacent cisoid amide bond
locked in that configuration by the two methyl groups flanking
its carbonyl group. The three stereocenters situated next to the
carbonyl and thiazoline moieties should not be ignored, for they
could epimerize at certain stages of the growth of the molecule
or at the end of the synthesis. Also, the architectural uncertainties
left unanswered by the structural elucidation studies, particularly
at C-3, C-4, and C-7,6demanded flexible strategies for the
construction of key intermediates, especially of the hydroxyl
decanoic acid fragment carrying the remote methoxy group of
halipeptins A (1a) and D (1d). It was with these considerations
in mind that we entered into the endeavor we are about to
describe, beginning with the retrosynthetic analysis of the
Retrosynthetic Analysis. Mindful of the known methods7
of converting an oxazoline to a thiazoline through the corre-
sponding hydroxymethyl thioamide (i.e., AfBfC, Figure 2),
we initially considered the possibility of obtaining the final
product (e.g., halipeptin D, 1d) from its oxazoline counterpart
(6) (a) In their total synthesis of halipeptin A, Ma et al. have shown the decanoic
acid fragment to be 3(S),4(R),7(S): Yu, S.; Pan, X.; Lin, X.; Ma, D. Angew.
Chem., Int. Ed. 2005, 44, 135-138. (b) Hara, S.; Makino, K.; Hamada, Y.
Tetrahedron Lett. 2006, 47, 1081-1085.
(7) (a) Wipf, P.; Miller, C. P.; Venkatraman, S.; Fritch, P. C. Tetrahedron
Lett. 1995, 36, 6395-6398. (b) Wipf, P.; Fritch, P. C. Tetrahedron Lett.
1994, 35, 5397-5400. (c) Lafargue, P.; Guenot, P.; Lellouche, J.-P.
Heterocycles 1995, 41, 947-958.
Figure 1. Structures of halipeptins A-D and analogues thereof.
Figure 2. Oxazoline-thiazoline conversion.
Total Synthesis of Halipeptins A and D
A R T I C L E S
J. AM. CHEM. SOC. 9 VOL. 128, NO. 13, 2006 4461
as shown in Figure 2. Such an approach would not only allow
easier entry into the larger (and therefore less strained) mac-
rocycle of the required hydroxyl amide precursors, but would
also give birth to a new series of designed molecules for testing,
the oxazoline halipeptins. Our investigations along these lines
have been reported elsewhere, and although they provided entry
into a variety of oxazoline analogues of the halipeptins, they
failed to deliver the natural products.3
A new strategy for the total synthesis of halipeptins A (1a)
and D (1d) was, therefore, sought along the path depicted retro-
synthetically in Figure 3.5Thus, retrolactamization of 1a and
1d, accompanied by functional group manipulations, led to azide
methyl esters 7a and 7d as potential open-chain precursors
(Figure 3). Subsequent dismantling of the thiazoline ring allowed
the generation of hydroxy thioamides 8a and 8d, while dis-
connection at the indicated amide bond led to the hydroxy amino
esters 9a and 9d and carboxy azido ester 12 as potential building
blocks for these constructions (Figure 3). Further disconnection
of 9a and 9d revealed L-alanine derivative 10 and dipeptide
equivalents 11a and 11d, the latter two intermediates being
traced to the simpler starting materials 13, and 14a and 14d. A
similar simplification of ester 12 as shown led to hydroxy acid
15 and D-alanine derivative 16 as the starting points for that
construction (Figure 3). The strategy thus derived from this
analysis had the advantages of convergence and late-stage
installment of the thiazoline moiety thought to be necessary for
avoiding early problems of epimerization of the center R to it.8
Construction of Building Blocks. Following our first synthe-
sis3of the hydroxy decanoic acid fragment 15,6,9we developed
a second generation synthesis to this building block starting from
commercially available bromo epoxide 1710as shown in Scheme
1. Thus, regioselective opening of the epoxide ring within 17
with EtMgBr in the presence of CuI gave the corresponding
bromo alcohol in 99% yield, whose exposure to TBSOTf and
2,6-lutidine furnished TBS-protected compound 18 (99% yield).
Displacement of the bromide residue from the latter compound
(18) with NaI resulted in the quantitative formation of the
desired iodide 19, which was used to alkylate the Myers
auxiliary 24 (nBuLi-iPr2NH, LiCl, THF, -78 f 25 °C),11
(8) (a) Hirotsu, Y.; Shiba, T.; Kaneko, T. Bull. Chem. Soc. Jpn. 1970, 43,
1870-1873. (b) Wipf, P.; Fritch, P. C. Tetrahedron Lett. 1994, 35, 5397-
(9) For alternative syntheses of this decanoic acid, see: Della Monica, C.;
Maulucci, N.; De Riccardis, F.; Izzo, I. Tetrahedron: Asymmetry 2003,
14, 3371-3378 and ref 6.
(10) (R)-4-Bromo-1,2-epoxybutane (17) is commercially available from “Synthon
Chiragenics Corporation”, 7 Deer Park Drive, Monmouth Junction, NJ
Figure 3. Second generation retrosynthetic analysis of halipeptins A and D (1a, 1d).
A R T I C L E S
Nicolaou et al.
4462 J. AM. CHEM. SOC.9VOL. 128, NO. 13, 2006
leading to hydroxy amide 20 in 66% yield and 12:1 diastereo-
selectivity (determined by1H NMR spectroscopic analysis).
Treatment of 20 with TBAF, followed by methylation (NaH,
MeI) of the free hydroxyl groups within the resulting diol (21,
99% yield), afforded the corresponding bis-methoxy compound,
which was cleaved to afford alcohol 22 (nBuLi-iPr2NH, BH3‚
NH3, 82% yield). The latter compound (22) was oxidized under
Swern conditions ((COCl)2-DMSO, Et3N), leading to the
corresponding aldehyde in good yield. Mukaiyama aldol reaction
of the resulting aldehyde in the presence of the boron compound
25 yielded hydroxy ester 23 in 89% yield and ca. 95:5
diastereoselectivity.12Finally, LiOH-induced ester hydrolysis
of 23 led to the targeted decanoic hydroxy acid derivative 15
in quantitative yield. This 10-step synthesis of 15 (44% overall
yield) is our preferred route to this intermediate.
The next intermediate to be targeted was the hydroxyisoleu-
cine derivative 14a (Scheme 2).6,13Its construction began with
homopropargylic alcohol 26, which was protected as a TBDPS
ether (TBDPSCl, imid., 99% yield) and then converted to pro-
pargyl alcohol derivative 28 by reaction with nBuLi and p-
formaldehyde (72% yield). The acetylenic bond within com-
pound 28 was then selectively reduced to afford cis olefin 29
through the action of hydrogen in the presence of 5% Pd/BaSO4
and quinoline (98% yield).14Sharpless asymmetric epoxidation
[Ti(iPrO)4, L-DET, tBuOOH] of the latter compound (29) gave
hydroxy epoxide 30 in 99% yield and 92% ee.14(This enan-
tioriched compound was curried through until the coupling with
(11) (a) Myers, A. G.; Yang, B. H.; Chen, H.; Gleason, J. L. J. Am. Chem. Soc.
1994, 116, 9361-9362. (b) Myers, A. G.; Yang, B. H.; Chen, H.;
McKinstry, L.; Kopecky, D. J.; Gleason, J. L. J. Am. Chem. Soc. 1997,
(12) Kiyooka, S.-I.; Kaneko, Y.; Komura, M.; Matsuo, H.; Nakano, M. J. Org.
Chem. 1991, 56, 2276-2278.
(13) For other syntheses of this building block, see: (a) Izzo, I.; Avallone, E.;
Della Corte, L.; Maulucci, N.; De Riccardis, F. Tetrahedron: Asymmetry
2004, 15, 1181-1186. (b) Hara, S.; Makino, K.; Hamada, Y. Tetrahedron
2004, 60, 8031-8035.
(14) For these steps, see: Still, W. C.; Ohmizu, H. J. Org. Chem. 1981, 46,
Scheme 1. Second Generation Synthesis of Hydroxydecanoic
Acid Building Block (15)a
a(a) EtMgBr (1.0 M in THF, 2.0 equiv), CuI (1.1 equiv), THF, 0 °C, 30
min, 99%; (b) TBSOTf (1.5 equiv), 2,6-lutidine (2.0 equiv), ether, 0 °C,
15 min, 99%; (c) NaI (2.0 equiv), acetone, 25 °C, 12 h, 99%; (d) (i) nBuLi
(2.5 M in hexanes, 4.0 equiv), LiCl (10.0 equiv), iPr2NH (4.2 equiv), THF,
-78 °C, 15 min; (ii) 24 (2.0 equiv) -78 f 25 °C, 19 h, 66% (12:1); (e)
TBAF (1.0 M in THF, 4.0 equiv), THF, 25 °C, 4 h, 99%; (f) NaH (60%
suspension in oil, 3.0 equiv), MeI (4.0 equiv), THF, 25 °C, 12 h, 95%; (g)
nBuLi (2.5 M in hexanes, 4.0 equiv), iPr2NH (4.0 equiv), BH3‚NH3(4.0
equiv), THF, -78 f -25 °C, 16 h, 82%; (h) (COCl)2(2.0 equiv), DMSO
(3.0 equiv), CH2Cl2, -78 °C; then Et3N (4.0 equiv), -78 °C, 3 h; (i) 25
(1.0 equiv), Me2CdC(OMe)OTMS (1.2 equiv), CH2Cl2, -78 °C, 1.5 h,
89%; (j) LiOH‚H2O (6.0 equiv), MeOH/H2O (4:1), 25 °C, 20 h, 99%. TBAF
) tetra-n-butylammonium fluoride; TBS ) tert-butylsilyl; TMS ) tri-
methylsilyl.; Tf ) trifluoromethanesulfonyl.
Scheme 2. Construction of TBDPS-Protected
δ-Hydroxy-L-isoleucine Methyl Ester (14a)a
a(a) TBDPSCl (1.1 equiv), imidazole (2.2 equiv), CH2Cl2, 25 °C, 16 h,
99%; (b) nBuLi (2.5 M in hexanes, 1.05 equiv), p-formaldehyde (1.1 equiv),
ether, -78 f 25 °C, 14 h, 72%; (c) H2(1 atm), 5% Pd/BaSO4(7 wt %),
quinoline (5 wt %), MeOH, 25 °C, 20 min, 98%; (d) Ti(OiPr)4(0.3 equiv),
L-DET (0.35 equiv), tBuOOH (2.0 equiv), MS (4 Å), CH2Cl2, 20 °C, 17 h,
99%, 92% ee (Mosher ester); (e) NaIO4(4.1 equiv), RuCl3‚H2O (2.2 mol
%), CCl4/CH3CN/H2O (1:1:1.5), 25 °C, 2 h, 80%; (f) AlMe3(3.0 equiv),
hexanes, 25 °C, 30 h, 80%; (g) MeI (1.3 equiv), K2CO3(1.1 equiv), acetone,
25 °C, 18 h, 90%; (h) Tf2O (1.5 equiv), pyridine (10.0 equiv), CH2Cl2, 0
°C, 20 min; (i) NaN3(1.5 equiv), DMF, 25 °C, 40 min, 75% (two steps);
(j) H2(1 atm), 10% Pd/C (30 wt %), EtOH, 25 °C, 2.5 h, 99%. DET )
diethyl tartrate; DMF ) N,N-dimethylformamide; MS ) molecular sieves;
TBDPS ) tert-butyldiphenylsilyl; Tf ) trifluoromethanesulfonyl.
Scheme 3. Preparation of D-Alanine Equivalent 37a
a(a) TfN3(3.0 equiv), CuSO4‚5H2O (0.5 equiv), Et3N (4.0 equiv), MeOH,
H2O, CH2Cl2, 25 °C, 16 h, 80%; (b) (COCl)2(1.0 equiv), DMF, 25 °C, 30
min. DMF ) N,N-dimethylformamide; Tf ) trifluoromethanesulfonyl.
Total Synthesis of Halipeptins A and D
A R T I C L E S
J. AM. CHEM. SOC. 9 VOL. 128, NO. 13, 2006 4463
enantiomerically pure 39a at which point the pure diastereomer
40a was separated by chromatography, see below). Oxidation
of hydroxy epoxide 30 with NaIO4in the presence of catalytic
amounts of RuCl3furnished carboxylic acid 31 whose reaction
with AlMe3proved both regioselective and efficient, leading
to hydroxy acid 32 (80% yield) into which the required methyl
group had been incorporated in a stereoselective manner.14
Treatment of hydroxy acid 32 with near stoichiometric amounts
of K2CO3and MeI resulted in the formation of hydroxy methyl
ester 33 in 90% yield. Triflate formation (Tf2O, py) followed
by reaction of the resulting triflate (34) with NaN3led to azide
35 in 75% overall yield for the two steps. With its inverted
stereochemistry, azide ester 35 served well as a precursor to
the desired amino ester 14a upon reduction with hydrogen in
the presence of 10% Pd/C (99% yield).
Acid chloride 37, an L-alanine equivalent, was prepared by
the method of Wong et al.15as shown in Scheme 3. Thus,
reaction of L-alanine (36) with TfN3in the presence of CuSO4‚
6H2O and Et3N resulted in the formation of azide 16 (with
retention of stereochemistry) in 80% yield. Treatment of the
latter compound (16) with (COCl)2 in DMF furnished acid
chloride 37 in high yield, an acylating agent used directly and
without purification in the subsequent coupling (vide infra).
Assembly of Building Blocks into Halipeptins A and D.
We begin the description of the building block assembly with
the construction of alanine-R-methylserine-N-methylisoleucine
(15) Formed in situ by reaction of oxalyl chloride with the corresponding acid:
(a) Alper, P. B.; Hung, S.-C.; Wong, C.-H. Tetrahedron Lett. 1996, 37,
6029-6032. (b) Lundquist, J. T.; Pelletier, J. C. Org. Lett. 2001, 3, 781-
Scheme 4. Synthesis of Thioamides 9a and 9da
a(a) TBSOTf (2.0 equiv), 2,6-lutidine (4.0 equiv), CH2Cl2, 0 °C, 1.5 h, 98%; (b) LiOH‚H2O (3.0 equiv), MeOH/H2O (4:1), 0 f 25 °C, 2 h, 100%; (c)
14a (1.0 equiv), EDC (1.2 equiv), HOAt (1.2 equiv), iPr2NEt (3.0 equiv), CH2Cl2, 25 °C, 17 h, 55% (38a); 14d (2.0 equiv), EDC (2.0 equiv), HOAt (2.0
equiv), iPr2NEt (3.0 equiv), CH2Cl2, 25 °C, 17 h, 91% (40d); (d) TESOTf (1.3 equiv), 2,6-lutidine (2.0 equiv), CH2Cl2, 0 °C, 20 min, 85%; (e) NaH (4.0
equiv), MeOTf (2.5 equiv), THF, 0 °C, 1 h, 83% (41a); NaH (3.0 equiv), MeI (4.0 equiv), DMF, 0 °C, 1 h, 96% (41d); (f) aqueous HCl (1 N; 20 equiv),
THF, 25 °C, 1 h, 100%; (g) TBAF (3.0 equiv), THF, 25 °C, 2 h, 98%; (h) EDC (4.0 equiv), Cbz-L-Ala-OH (4.0 equiv), CH2Cl2, 25 °C, 0.5 h; then 43a or
43d, 18 h, 44a, 88%; 44d, 96%; (i) PMe3(1.8 equiv), toluene, 25 °C, 2 h, 47a, 79%; 47d, 83%; (j) H2(1 atm), 20% Pd(OH)2/C (30 wt %), EtOH, 25 °C,
1 h, 100%; (k) H2S (excess), MeOH/Et3N (2:1), 25 °C, 48-72 h, complete conversion, product not isolated. Cbz ) benzyloxycarbonyl; CDI )
carbonyldiimidazole; EDC ) 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride; HOAt ) 1-hydroxy-7-azabenzotriazole; TBAF ) tetra-n-
butylammonium fluoride; TBS ) tert-butyldimethylsilyl; TES ) triethylsilyl.
A R T I C L E S
Nicolaou et al.
4464 J. AM. CHEM. SOC.9VOL. 128, NO. 13, 2006
tripeptides 44a and 44d (Scheme 4). Our initial forays to these
tripeptide fragments were met with failure due to the propensity
of the intermediate dipeptides to undergo intramolecular bridging
to diketopiperazines. We, therefore, adopted an alternative
approach to these tripeptide units that did not involve a free
amino group as part of the growing molecule as depicted in
Scheme 4. Thus, the previously known R-hydroxy-R-methyl-
serine equivalent 3816was silylated (TBSOTf, 2,6 lut., 98%
yield) and the resulting TBS ether methyl ester was hydrolyzed
(LiOH, 100% yield) to afford carboxylic acid 39a. This
R-methyl-serine equivalent was then coupled to the δ-hydroxy
isoleucine methyl ester derivative 14a6,13(EDC, HOAt, 55%
yield) to furnish dipeptide 40a. TES protection (TESOTf, 2,6
lut., 85% yield) of the hydroxyl group of the latter compound
(40a) allowed N-methylation (NaH, MeOTf, 83% yield) of its
amide moiety through the intermediacy of the bis-silyl ether
41a to afford compound 42a. Exposure of the latter substance
to 1 N aqueous HCl in THF then led to selective mono-
desilylation of the latter compound (42a) to give the desired
hydroxy azido fragment 43a required for the total synthesis of
halipeptin A (1a). The hydroxy azido fragment 43d required
for the construction of halipeptin D (1d) was prepared by TBAF-
induced desilylation (98% yield) of the TBS ether 42d, whose
synthesis has been described elsewhere.3
Each of the hydroxy azido dipeptides thus obtained (43a and
43d) was converted to the corresponding ester by coupling with
Cbz-L-Ala-OH (EDC, 44a, 88% yield; 44d, 96% yield) as a pre-
lude to the tripeptide construction. It was anticipated that reac-
tion of the azide group with a phosphine would result in aza-
ylid formation (i.e., 45a and 45d) and subsequent addition to
the nearest grouping, an occurrence that was expected to lead
to the desired oxazolines (47a and 47d) by expulsion of phos-
phine oxide from the corresponding four-membered ring phos-
phine derivatives (i.e., 46a and 46d) as shown in Scheme 4.17
Much to our delight, this cascade sequence was realized in
the laboratory by treating 44a or 44d with PMe3in benzene or
toluene solution at room temperature. The coveted oxazo-
lines 47a and 47d were obtained in 79% and 83% yield, respect-
ively, accompanied only by small amounts of the undesired
dihydropyrazinones 48a (17%) and 48d (16%), respectively.
The latter compounds were presumably formed upon hydrolysis
(16) (a) Avenoza, A.; Cativiela, C.; Corzana, F.; Peregrina, J. M.; Sucunza, D.;
Zurbano, M. M. Tetrahedron: Asymmetry 2001, 12, 949-958. (b) Smith,
N. D.; Goodman, M. Org. Lett. 2003, 5, 1035-1038.
(17) For previous examples of oxazoline construction through aza-Wittig
cyclization, see: (a) Gololobov, Yu. G.; Gusar, N. I.; Chaus, M. P.
Tetrahedron 1985, 41, 793-799. (b) Mulzer, J.; Meier, A.; Buschmann,
J.; Luger, P. Synthesis 1996, 123-132. (c) Kato, H.; Ohmori, K.; Suzuki,
K. Synlett 2001, 1003-1005. For thiazoline formations with triphenylphos-
phine, see: (d) Kok, G. B.; Campbell, M.; Mackey, B.; von Itzstein, M. J.
Chem. Soc., Perkin Trans. 1 1996, 23, 2811-2815. (e) Chen, J.; Forsyth,
C. J. Org. Lett. 2003, 5, 1281-1284. (f) Chen, J.; Forsyth, C. J. J. Am.
Chem. Soc. 2003, 125, 8734-8735.
Scheme 5. Synthesis of epi-isoleucine Halipeptin D Stereoisomer 4da
a(a) 37 (10.0 equiv), 4-DMAP (0.5 equiv), Et3N (12 equiv), DMF, 50 °C, 2 h, 94%; (b) 9d (2.0 equiv), PyAOP (2.5 equiv), iPr2NEt (2.5 equiv), DMF,
25 °C, 17 h, 92%; (c) TBSOTf (1.2 equiv), 2,6-lutidine (1.5 equiv), CH2Cl2, 0 °C, 0.5 h, 99%; (d) Me3SnOH (20 equiv), 1,2-dichloroethane, 80 °C, 48 h,
52% at 79% conversion; (e) PMe3(3.0 equiv), THF/H2O (9:1), 25 °C, 48 h; (f) HATU (2.0 equiv), HOAt (2.0 equiv), iPr2NEt (2.0 equiv), CH2Cl2, 25 °C,
18 h, 38% (two steps); (g) 50% aqueous HF/CH3CN (1:99 by volume), 25 °C, 24 h, 100%; (h) DAST (2.0 equiv), CH2Cl2, -78 f 20 °C, 2 h, 80%. DAST
) diethylaminosulfurtrifluoride; 4-DMAP ) 4-(dimethylamino)pyridine; DMF ) N,N-dimethylformamide; HATU ) O-(7-azabenzotriazol-1-yl)-N,N,N′,N′-
tetramethyluronium hexafluorophosphate; HOAt ) 1-hydroxy-7-azabenzotriazole; PyAOP ) (7-azabenzotriazol-1-yloxy) tripyrrolidinophosphonium
hexafluorophosphate; TBS ) tert-butyldimethylsilyl.
Total Synthesis of Halipeptins A and D
A R T I C L E S
J. AM. CHEM. SOC. 9 VOL. 128, NO. 13, 2006 4465
of the aza-ylids by traces of water leading to the corresponding
amines, which then engaged the proximal methyl ester moieties
in ring formation. Chromatographic purification of 47a and 47d
then allowed their efficient conversion to thioamides 9a and
9d, respectively, by a two-step sequence involving hydrogenoly-
sis of the Cbz group [H2, 10% Pd(OH)2/C, 100% yield] and sub-
sequent thiolysis of their oxazoline rings (H2S, Et3N)7aas shown
in Scheme 4. The obtained crude products 9a and 9d were
employed in the subsequent step without further purification.
Our first attempt to synthesize halipeptin D (1a) led, rather
unexpectedly, to epi-isoleucine halipeptin D (4d) as shown in
Figure 4. Selected nOe correlations for halipeptin A and D analogues 3a, 3d, and 4a, 4d.
Scheme 6. Completion of the Total Synthesis of Halipeptins A (1a) and D (1d) and Their Analogues 3a, 3d, 4a, and 4da
a(a) 9a (1.0 equiv), PyAOP (1.5 equiv), iPr2NEt (2.0 equiv), DMF, 25 °C, 17 h, 71%; (b) DAST (1.5 equiv), CH2Cl2, -78 f 20 °C, 1 h, 53a, 85%; 53d,
84%; (c) Me3SnOH (20 equiv), 1,2-dichloroethane, 90 °C, 48-72 h, 82% from 53a; 85% from 53d; (d) PMe3(2.5 equiv), THF/H2O (9:1), 25 °C, 2 h; (e)
HATU (2.0 equiv), HOAt (2.0 equiv), K2CO3(10.0 equiv), CH2Cl2, 25 °C, 24 h, 55, 31%; 56, 10%; 57, 14%; 1d, 25%; 3d, 5%; 4d, 18%; (f) TAS-F (5.0
equiv), DMF, 25 °C, 24 h, 1a, 85%; 3a, 81%; 4a, 85%. DAST ) diethylaminosulfurtrifluoride; DMF ) N,N-dimethylformamide; HATU )
O-(7-azabenzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophosphate; HOAt ) 1-hydroxy-7-azabenzotriazole; PyAOP ) (7-azabenzotriazol-1-
yloxy) tripyrrolidinophosphonium hexafluorophosphate; TBDPS ) tert-butyldiphenylsilyl; TAS-F ) tris-(dimethylamino)sulfur-(trimethylsilyl)difluoride;
Tf ) trifluoromethanesulfonyl.
A R T I C L E S
Nicolaou et al.
4466 J. AM. CHEM. SOC.9VOL. 128, NO. 13, 2006
Fearing epimerization of the stereocenter adjacent to the
thiazoline moiety, we followed a strategy that called for late
stage construction of the thiazoline ring.8Thus, treatment of
hydroxy acid fragment 15 with excess azido chloride 37
(prepared in situ from the corresponding acid and oxalyl
chloride)15in the presence of Et3N and 4-DMAP yielded the
sterically crowded ester 12 in 94% yield. That no epimerization
had occurred during this coupling reaction at the azide-bearing
site, despite the basic conditions employed, was confirmed by
synthesizing the corresponding epimeric material through a
similar sequence utilizing 15 and ent-37.
With both fragments 12 and 9d in hand, we were in a position
to assemble the entire chain needed for the macrocycle
construction. The coupling of these two fragments was best
achieved through the action of PyAOP in the presence of
iPr2NEt, leading to advanced intermediate 8d in 92% yield.
Protection of the hydroxyl group (TBSOTf, 2,6 lut., 99% yield)
within the latter intermediate then produced TBS ether 49d,
whose Me3SnOH-induced methyl ester cleavage led to car-
boxylic acid azide 50d in 52% yield (79% conversion).18The
following two steps were carried out without purification of
the intermediate amino acid. Thus, the Staudinger reduction of
carboxylic acid azide 50d with PMe3in THF:H2O (9:1) was
followed by macrolactamization employing HATU, HOAt, and
iPr2NEt in CH2Cl2(ambient temperature) to afford macrocycle
51d in 38% overall yield for the two steps. Removal of the
TBS group from the latter compound (51d) was smoothly and
quantitatively achieved by exposure to 50% aqueous HF in CH3-
CN, leading to hydroxy thioamide macrocycle 52d. However,
(18) (a) Nicolaou, K. C.; Estrada, A. A.; Zak, M.; Lee, S. H.; Safina, B. S.
Angew. Chem., Int. Ed. 2005, 44, 1378-1382. (b) Furlan, R. L. E.; Mata,
E. G.; Mascaretti, O. A. J. Chem. Soc., Perkin Trans. 1 1998, 355-358.
(c) Furlan, R. L. E.; Mata, E. G.; Mascaretti, O. A.; Pena, C.; Coba, M. P.
Tetrahedron 1998, 54, 13023-13034. (d) Furlan, R. L. E.; Mata, E. G.;
Mascaretti, O. A. Tetrahedron Lett. 1996, 37, 5229-5232.
Scheme 7. Incorporation of D-Alanine into the Total Synthesis of Halipeptin 3da
a(a) CDI (4.0 equiv), Cbz-D-Ala-OH (4.0 equiv), CH2Cl2, 25 °C, 18 h, 69%; (b) PMe3 (1.8 equiv), toluene, 25 °C, 2 h, 80%; (c) H2 (1 atm), 10%
Pd(OH)2/C (30 wt %), EtOH, 25 °C, 1 h, 99%; (d) H2S (excess), MeOH/Et3N (2:1), 25 °C, 72 h; (e) 43 (1.0 equiv), PyAOP (1.5 equiv), iPr2NEt (2.0 equiv),
DMF, 25 °C, 18 h, 63%; (f) DAST (1.5 equiv), CH2Cl2, -78 f 20 °C, 1 h, 72%; (g) Me3SnOH (20 equiv), 1,2-dichloroethane, 90 °C, 72 h; (h) PMe3(2.5
equiv), THF/H2O (9:1), 25 °C, 2 h; (i) HATU (2.0 equiv), HOAt (2.0 equiv), iPr2NEt (6.0 equiv), CH2Cl2, 25 °C, 24 h, 59%. Cbz ) benzyloxycarbonyl;
CDI ) carbonyldiimidazole; DAST ) diethylaminosulfurtrifluoride; DMF ) N,N-dimethylformamide; EDC ) 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide
hydrochloride; HATU ) O-(7-azabenzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophosphate; HOAt ) 1-hydroxy-7-azabenzotriazole; PyAOP
) (7-azabenzotriazol-1-yloxy) tripyrrolidinophosphonium hexafluorophosphate.
Table 1. Cytotoxicity (IC50, µM)aof Synthetic Halipeptins A (1a) and D (1d) and Their Analogues (3a, 3d, 4a, 4d, 5d, and 6d)
1a1d3a3d 4a 4d5d 6d
aThe concentration at which the compound inhibits 50% of the cell growth.bHuman colon carcinoma cell line.
Total Synthesis of Halipeptins A and D
A R T I C L E S
J. AM. CHEM. SOC. 9 VOL. 128, NO. 13, 2006 4467
although the performance of DAST in the final step to prepare
the targeted halipeptin D (1d) was admirable in terms of
furnishing (80% yield) the desired thiazoline moiety,19this route
failed in reaching its goal, yielding instead epi-isoleucine
halipeptin D (4d) as indicated in Scheme 5. Apparently,
somewhere along the line, most likely at the macrocyclization
step, the conditions employed caused epimerization at the
indicated (*) site, a fact supported by NMR spectroscopic studies
of 4d (1H,
consequence of the sensitivity of these molecules toward
epimerization, forced upon us a change in strategy and tactics
toward halipeptins D (1d) and A (1a), as will be described
Our revised strategy for the total synthesis of halipeptins A
(1a) and D (1d) and their analogues improvising for thiazoline
construction prior to macrocyclization is depicted in Scheme
6. Thus, coupling of the sterically congested ester 12 with amino
fragment 9a or 9d (see Scheme 4 for their preparation) under
the influence of PyAOP and iPr2NEt led to thiamides 8a (71%
yield) and 8d (92% yield), respectively. Reaction of these
intermediates with DAST in CH2Cl2 at -78 °C resulted in
smooth thiazoline formation19(53a, 82% yield; 53d, 85% yield).
The methyl ester group of the resulting compounds (53a, 53d)
13C, ROESY, see Figure 4). This outcome, a
was then cleaved by treatment with Me3SnOH, leading to the
corresponding carboxylic acid azides (82%, 85%, respectively),18
which were reduced under Staudinger conditions [PMe3, THF:
H2O (9:1)] to obtain the crude amino acids 54a and 54d, ready
for macrocyclization. Our first attempts to cyclize the amino
acids under the same conditions as before (i.e., HATU, HOAt,
iPr2NEt) led to the formation of the same epimerized halipeptin
D (4d) in 17% overall yield from the corresponding azido
carboxylic acid. This time, however, 4d was accompanied by
halipeptin D (1d) and analogue 3d (epimer at the indicated (*)
site adjacent to the thiazoline ring, ROESY analysis, see Figure
4), which were formed in 13% and 5% yield, respectively, from
the corresponding azido carboxylic acid. Switching the base
from iPr2NEt to K2CO3in the macrolactamization step resulted
in an increase of both the yield and the preference for halipeptin
D (1d) to 25%, with 3d and 4d being formed in 5% and 18%
yield, respectively. These isomeric halipeptins D could easily
be separated by preparative thin-layer chromatography (PTLC,
silica gel) due to the rather large differences in their Rfvalues.
In a similar way, the HATU/HOAt/K2CO3macrolactamiza-
tion protocol was employed to produce the TBDPS derivative
of halipeptin A (55) from amino acid 54a in 31% yield, together
with its epimers 56 (10% yield) and 57 (14% yield) after
separation with PTLC (Scheme 6). Treatment of these deriva-
tives with TAS-F generated halipeptin D (1d, 86% yield) and
(19) Lafargue, P.; Guenot, P.; Lellouche, J.-P. Heterocycles 1995, 41, 947-
Figure 5. Dose-dependent inhibition on the first (0-6 h) and second phases (24-96 h) of carrageenan-induced paw edema by synthetic halipeptins (1a, 1d,
3a, 3d, 5d, and 6d). Data are expressed as mean ( s.e. mean and analyzed by using ANOVA for multiple comparisons followed by Dunnett’s test. Statistical
significance is set at p < 0.05.
A R T I C L E S
Nicolaou et al.
4468 J. AM. CHEM. SOC.9VOL. 128, NO. 13, 2006
its epimers 3a (81% yield, ROESY analysis, see Figure 4) and
4a (85% yield, ROESY analysis, see Figure 4).
At this juncture, we became curious whether the unnatural
halipeptin D 3d containing the thiazoline-bound D-alanine
residue was more stable than the naturally occurring substance
halipeptin D (1d) and, therefore, set out to synthesize it. Scheme
7 summarizes this endeavor, which proceeded along the same
lines as before and in similar yields, except for the macrolac-
tamization step that yielded only the unnatural halipeptin D (3d),
and in good yield (59% yield, two steps). The absence of any
epimerized products at the end of this synthesis reflects the
higher stability of this molecule (3d) relative to its natural isomer
(1d) and ensures a relatively easy access to this unnatural
Biological Evaluations of Synthetic Halipeptins A and
D and Analogues
Cytotoxic Properties. As mentioned above, Faulkner and
Manam reported that halipeptin D (1d) exhibited potent cyto-
toxic properties [IC50) 7 nM against human colon cancer HCT-
116 cell line and an average IC50value of 420 nM against a
BMS ODCP (oncology diverse cell panel) of tumor cell lines].4
These results were in sharp contrast to those of Gomez-Paloma
et al. for their halipeptins A-C (1a-1c), which apparently were
devoid of any significant cytotoxicity properties.1,2We have
also reported elsewhere that the oxazoline analogues of hal-
ipeptin D, compounds 5d and 6d (Figure 1), were found to have
only weak cytotoxicity.3Intrigued by the previous findings and
by the discrepancy between the results of the Gomez-Paloma
group and Faulkner and Manam, we proceeded to test synthetic
halipeptins A (1a) and D (1d), and their unnatural analogues,
3a, 3d, 4a, and 4d, for cytotoxic activity. As shown in Table 1,
all of these synthetic halipeptins exhibited only weak cytotox-
icity against tumor cells despite the previous reports regarding
naturally derived halipeptin D (1d) by Faulkner and Manam.3,4
These observations may be explained if we assume that naturally
derived halipeptin D (1d) was contaminated by one or more
potent contaminants, presumably coexisting within the marine
species. Thus, we believe that the halipeptins are devoid of
significant cytotoxicity, a fact that makes their anti-inflammatory
properties even more meaningful and potentially useful.
Anti-inflammatory Properties. It was previously demon-
strated that natural halipeptin A (1a) exhibited 60% reduction
of carrageenan-induced paw edema in mice at the intraperitoneal
dose of 0.3 mg kg-1body weight.1,2The significance of this
activity can be better appreciated when compared to those of
the commercial agents, indomethacin and naproxen, which
exhibit 40-fold and 130-fold lower potencies in the same test,
respectively, than halipeptin A (1a) (ED50 12 mg kg-1for
indomethacin, and 40 mg kg-1for naproxen).20These findings
prompted us to investigate the anti-inflammatory properties of
our synthetic halipeptins (1a, 1d, 3a, 3d, 5d,3and 6d,3Figure
1) using the same biological assay. As shown in Figure 5,
administration of halipeptin A (1a, 1 mg kg-1) resulted in
reduction of the volume in the second phase of the mouse paw
edema. More specifically, a significant inhibition occurred at
24 h (*p < 0.05), 48 h, and 72 h (***p < 0.001) as compared
to the corresponding control time points. These results confirmed
the original findings of the Gomez-Paloma group1,2regarding
the anti-inflammatory activity of natural halipeptin A (1a).
Synthetic halipeptin D (1d) exhibited an anti-inflammatory
(20) Calhoun, W.; Chang, J.; Carlson, R. P. Agents Actions 1987, 21, 306-
Figure 6. Dose-dependent inhibition of phase I (0-6 h) and phase II (24-
96 h) of carrageenan-induced paw edema by synthetic halipeptins (1a, 1d,
3a, 3d, 5d, and 6d). Data are expressed as mean ( s.e. mean and analyzed
by using ANOVA for multiple comparisons followed by Dunnett’s test.
Statistical significance is set at p < 0.05.
Figure 7. Inhibition of carrageenan-induced paw edema by synthetic
halipeptins (1a, 1d, 3a, 3d, 5d, and 6d). Data are expressed as area under
the curve (AUC) for phase I (A) and phase II (B) edema.
Total Synthesis of Halipeptins A and D
A R T I C L E S
J. AM. CHEM. SOC. 9 VOL. 128, NO. 13, 2006 4469
action similar to that observed for halipeptin A (1a) on the
second phase of the edema. In addition, halipeptin D (1d)
significantly reduced the edema at the 6 h time point of the
first phase (*p < 0.05), suggesting that its anti-inflammatory
action was faster on onset as compared to that of halipeptin A
(1a, Figure 5). The oxazoline analogue 6d displayed an anti-
inflammatory activity similar to that of halipeptin D (1d) in
the second phase of the mouse paw edema. However, oxazoline
6d displayed pronounced reduction of the mouse paw edema
in the first phase at the 6 h time point (Figure 5). Most
impressively, swelling was almost undetectable in the acute
phase (phase I) of the edema (within the first 6 h). This was
considerably more pronounced than those displayed by the
natural halipeptins A (1a) and D (1d).
Figure 6 depicts, in color, the graphs for phase I and phase
II of carrageenan-induced paw edema by the various synthetic
halipeptins for easy comparison purposes. As shown in that
figure, halipeptins 3a, 3d, and 5d did not exhibit significant
anti-inflammatory activity as compared to 1a, 1d, and 6d.
Figure 7 shows the total reduction of edema for phase I (graph
A) and phase II (graph B) as measured by the under the curve
area (AUC) for synthetic halipeptins 1d, 5d, 1a, 6d, 3d, and
3a, demonstrating the potent early (phase I) effects of halipeptin
D (1d) and its oxazoline counterpart (6d; the AUC for both
compounds was significantly reduced as compared to vehicle
### p < 0.001) and the late effect (phase II) of halipeptin A
(1a) (### p < 0.001).
The potent overall effect of halipeptins A (1a), D (1d), and
epi-oxazoline halipeptin D (6d), as measured by the area under
the curve (AUC), is demonstrated in Figure 8.
The described chemistry provides entries into the halipeptin
family of compounds including the naturally occurring halipep-
tins A (1a) and D (1d), and their epimers 3a, 3d and 4a, 4d.
By rendering these materials available, as well as their oxazoline
analogues 5d3and 6d,3chemical synthesis allowed their
biological evaluation in cytotoxicity and inflammation assays.
The former investigations revealed no significant cytotoxicity
for any of the tested synthetic halipeptins, including the naturally
occurring halipeptins A (in line with previous reports)1,2and D
(in contrast to previous reports).4To explain the discrepancy
between synthetic and naturally derived halipeptin D (1d), we
assumed contamination of the natural sample by potent cytotoxic
impurity(ies) that may have been present in the original host of
this natural product.
In contrast to the cytotoxicity assays, our anti-inflammatory
tests involving carrageenan-derived mouse edema detected
strong anti-inflammatory activity for both synthetic halipeptins
A (1a, confirming the observation of Gomez-Paloma et al. with
a natural sample)1,2and D (1d), as well as of oxazoline analogue
6d.3The latter compound, in particular, exhibited remarkable
edema reduction in both the first and the second phases of the
condition, proving it to be a highly potent anti-inflammatory
agent. In view of these promising results, further studies in the
field may be warranted. These studies would include investiga-
tions directed toward the elucidation of the mechanism of action
of the halipeptins and the design and synthesis of analogues
that may prove useful as anti-inflammatory agents.
Acknowledgment. We thank Dr. D. H. Huang and Dr. G.
Siuzdak for NMR spectroscopic and mass spectrometric as-
sistance, respectively. Financial support for this work was
provided by the National Institutes of Health, the CaPCURE
Foundation, the Skaggs Institute for Chemical Biology, post-
doctoral fellowships from the Ernst Schering Research Founda-
tion (to D.S.) and the Royal Society-Fulbright Commission (to
D.A.L.), and a graduate fellowship from the Portuguese
Foundation for Science & Technology (to R.G.N.). R.G.N. is
an exchange graduate student to TSRI from REQUIMTECQFB,
Chemistry Department, FCT, Universidade Nova de Lisboa,
Supporting Information Available: Experimental procedures
and compound characterization, and full citation for ref 4b. This
material is available free of charge via the Internet at
Figure 8. Inhibition of carrageenan-induced paw edema by halipeptin A
(1a), halipeptin D (1d), and oxazoline analogue (6d). Data are expressed
as area under the curve (AUC) for combined phase I and phase II edema.
A R T I C L E S
Nicolaou et al.
4470 J. AM. CHEM. SOC.9VOL. 128, NO. 13, 2006