Synthesis and biological evaluation of (-)-laulimalide analogues

Article abstract of DOI:10.1016/j.bmcl.2003.12.001

Analogues of the marine natural product (-)-laulimalide were prepared by total synthesis and evaluated in vitro for anticancer activity.

Full text of DOI:10.1016/j.bmcl.2003.12.001

Bioorganic & Medicinal Chemistry Letters 14 (2004) 575–579
Synthesis and biological evaluation of (ꢀ)-laulimalide analogues
Brian M. Gallagher, Jr.,* Francis G. Fang, Charles W. Johannes,y Marc Pesant,
Martin R. Tremblay, Hongjuan Zhao, Kozo Akasaka, Xiang-yi Li, Junke Liu
and Bruce A. Littlefield
Eisai Research Institute, 100 Research Drive, Wilmington, MA 01887, USA
Received 15 October 2003; revised 21 November 2003; accepted 2 December 2003
Abstract—Analogues of the marine natural product (ꢀ)-laulimalide were prepared by total synthesis and evaluated in vitro for
anticancer activity.
# 2003 Elsevier Ltd. All rights reserved.
The marine natural product (ꢀ)-laulimalide (1, Fig. 1)
was first isolated in 1988 from several sources of marine
sponge and shown to be highly cytotoxic in a number of
human cancer cell lines.1 Despite the interesting and
challenging structural features of laulimalide, synthetic
efforts were relatively few until the report that 1 induces
microtubule polymerization and stabilization similar to
paclitaxel, but retains activity in a P-glycoprotein (PgP)
over-expressing multidrug resistant cell line.2 It was
recently reported that laulimalide binds to a different
site of the tubulin polymer than other known micro-
tubule stabilizers and retains activity against cell lines
containing mutations in the b-tubulin gene.3 These
reports have generated an enormous amount of excite-
ment about the therapeutic potential of this natural
product and as a result, there has been a surge of syn-
thetic efforts toward 1 resulting in total syntheses by
eight separate groups,4 as well as the syntheses of var-
ious fragments.5 Despite these synthetic efforts, the bio-
logical evaluation of only a few laulimalide analogues
have been reported to date.3,6
tionships. Herein we report the total syntheses and bio-
logical evaluation of 1 and some initial analogues.
The convergent strategy to synthesize 1 targets two
major fragments: a C.2–C.14 bottom half (3 or 4) and
C.15–C.27 upper half (2, Fig. 1). Indium mediated cou-
pling to form the C.14–C.15 bond allows access to both
diastereomers at C.15, as well as to the unnatural con-
figuration of the epoxide. Ring closure via Horner–
Wadsworth-Emmons (H-W-E) reaction4a from a C.3
aldehyde precursor (3) or Yamaguchi cyclization4b from
We began a program to explore the therapeutic poten-
tial of laulimalide and analogues. The initial goal was to
design a total synthesis capable of producing gram
quantities of 1 suitable for more in depth in vitro stud-
ies, and in vivo studies, with the flexibility to produce
analogues for exploration of the structure–activity rela-
* Corresponding author. Tel.: +978-661-7264; fax: +978-657-7715;
e-mail: brian_gallagher@eri.eisai.com
y
Current address: Infinity Pharmaceuticals, 780 Memorial Drive,
Cambridge, MA 02139, USA.
Figure 1. Synthetic strategy to access laulimalide.
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.12.001

576
B. M. Gallagher, Jr. et al. / Bioorg. Med. Chem. Lett. 14 (2004) 575–579
an alkynoic acid precursor (4) would allow for analo-
gues with a modified C.2–C.3 region. Intermediates 3
and 4 are both derived from (S)-citronellal, and inter-
mediate 2 is derived from d-arabinose and (R)-glycidol.
The enal 13 was constructed starting from the known
aldehyde 98 (Scheme 2), available in six steps from
d-arabinose. A number of different approaches to
attach the side chain were explored, but a modification
of the previously reported H-W-E reaction5l utilizing
phosphonate 11 proved to be the most efficient and
flexible. The known aldehyde 14 was most efficiently
prepared from R-glycidol.5f,g,j,n This approach pro-
duced tens of grams of the coupling partner enal 13.
Modification of the route described by Davidson5i
(Scheme 1) provided access to intermediate 7, which
could be converted to 3 for the H-W-E ring closing
strategy (not shown). However, the most efficient
approach to 1 involved incorporation of the C.2–C.3
alkyne in 4 prior to coupling with the top fragment.
This route produced tens of grams of the allyl bromide
4.
Indium mediated coupling of 4 with 13 produced aꢁ1:1
mixture at C.15 (Scheme 3). Macrolactonization4b of
alkynoic acid 15 proved to be the most efficient strategy
for the synthesis of 1. The C.15 diastereomers were
separated by chiral preparative HPLC (Chiralpak1
AD) following Lindlar reduction. Global deprotection
and Sharpless epoxidation provided 1.4e This strategy
proved capable of producing gram quantities of 1 for
evaluation and derivatization.
This route provided the requisite flexibility to produce
analogues of interest. Analogues lacking the epoxide
Scheme 1. Synthesis of the C.2–C.14 fragment.7 Reagents and condi-
tions: (a) (i) Ipc2BCH2CH=CH2, Et2O, ꢀ78 ꢂC; (ii) 3 M NaOH, 30%
H2O2, 75%; (b) 1-methoxy-1,2-propadiene, Pd(OAc)2, Et3N, CH3CN,
80 ꢂC, 65%; (c) Cl2(PCy3)2Ru=CHPh, CH2Cl2, 40 ꢂC, 85%; (d)
CH2CHOTBS, LiClO4, Et2O, 0 ꢂC; (e) NaBH4, THF/H2O (100/1),
0 ꢂC, 50%; (f) mCPBA, CH2Cl2, satd NaHCO3, 0 ꢂC; (g) 0.1 M H2SO4,
THF; (h) TBSCl, Et3N, DMAP, 0 ꢂC, 70%; (i) Pb(OAc)4, PhCH3,
95%; (j) (i) H2CN(CH3)2I, CH2Cl2; (ii) Et3N; (k) NaBH4, CeCl3,
MeOH, ꢀ78 ꢂC, 55%; (l) Ac2O, DMAP, pyridine, 95%; (m) H2SiF6,
CH3CN, 80%; (n) (COCl)2, DMSO, Et3N, CH2Cl2, ꢀ78 ꢂC, 80%; (o)
Ph3P, CBr4, CH2Cl2, ꢀ78 ꢂC, 70%; (p) n-BuLi, THF, ꢀ78 ꢂC, 80%;
(q) Ph3P, NBS, CH2Cl2, 0 ꢂC, 80%.
Scheme 3. Synthesis of laulimalide (1).7 Reagents and conditions: (a)
In, THF/H2O (3/1), HCl (cat.), 90%; (b) TBSOTf, pyridine, DMAP,
CH2Cl2, 90%; (c) (i) n-BuLi, THF, ꢀ78 ꢂC; (ii) CO2 (s); (d) DDQ,
CH2Cl2/pH 7 buffer (1/1); (e) 2,4,6-trichlorobenzoyl chloride, DMAP,
PhCH3, 50%; (f) H2, Lindlar cat., quinoline, hexane/CH2Cl2 (3/1),
95%; (g) chiral HPLC; (h) H2SiF6, CH3CN; (i) (+)-DIPT, Ti(OiPr)4,
t-BuOOH, CH2Cl2, ꢀ20 ꢂC, 80%.
Scheme 2. Synthesis of the C.15–C.27 fragment.7 Reagents and con-
ditions: (a) (i) (MeO)2POCH2CO2Me, NaH, THF, 0 ꢂC; (ii) Bu3P,
CH3CN, 60 ꢂC, 85%; (b) DIBAL, CH2Cl2, ꢀ78 ꢂC; (c) PivCl, pyridine,
CH2Cl2, 85%; (d) 1 M HCl/THF (1/1); (e) NaIO4, THF/H2O, 0 ꢂC; (f)
NaClO2, H2NSO3H, t-BuOH/H2O, 0 ꢂC; (g) TMSCHN2, PhCH3/
MeOH (3/1), 50%; (h) (i) (MeO)2POCH2Li, THF, ꢀ78 ꢂC; (ii) PivCl,
pyridine, CH2Cl2, 90%; (i) Et3N, LiCl, 14, THF, 0 ꢂC to rt, 70%; (j) l-
Selectride, THF, ꢀ78 ꢂC, 70%; (k) TBSOTf, imidazole, DMF, 0 ꢂC,
90%; (l) NaOMe, MeOH, 0 ꢂC, 80%; (m) (COCl)2, DMSO, Et3N,
CH2Cl2, ꢀ78 ꢂC, 85%.
Scheme 4. Synthesis of analogues 27–34. Reagents: (a) H2SiF6,
CH2Cl2/CH3CN/THF (2/2/1), 70%; (b) (+)-DIPT, Ti(OiPr)4, t-
BuOOH; (c) (i) (+)-DIPT, Ti(OiPr)4, t-BuOOH; (ii) RCOCl; (iii)
Et3N-3HF, CH3CN; (d) (i) (+)-DIPT, Ti(OiPr)4, t-BuOOH; (ii) NaH,
MeI; (iii) Et3N-3HF, CH3CN; (e) (i) Dess–Martin; (ii) Et3N-3HF,
CH3CN; (f) (i) Dess–Martin; (ii) MeONH3Cl; (iii) Et3N-3HF,
CH3CN.

B. M. Gallagher, Jr. et al. / Bioorg. Med. Chem. Lett. 14 (2004) 575–579
577
(e.g., 18 and 19) were prepared by deprotecting 16 and
the corresponding C.15 epimeric diastereomer. Fragment
3 was coupled to 13 in a sequence analogous to Scheme
3, which enabled the synthesis of analogues 20 and 21
via a H-W-E ring closure.4a The C.20 methoxy deriva-
tives 22–24 were synthesized by methylating inter-
mediate 12 under standard conditions (NaH, MeI).
Compound 24 was a minor by-product from the Lindlar
reduction. Acylation (Ac2O, pyridine) of 1 was reason-
ably selective for the C.20 alcohol to produce 25, but
also yielded some of the bis-acylated derivative 26. The
C.20 TBS ether (27) was synthesized as shown in
Scheme 4.
Analogues at the C.15 position (28–34) were synthesized
as shown in Scheme 4. Mono-deprotection of 16 occur-
red selectively at C.15 using H2SiF6. Following epox-
idation and/or derivatization of C.15, the C.20 TBS
ether was removed using Et3N-3HF.4i C.16-C.17 epi-
meric epoxide analogues (35 and 36) were made in an
Table 1. In vitro growth inhibition activities of laulimalide (1) and analogues 18–36
Compd
R1
R2
H
R3
H
X
Y
MDA-MB-435a
HT-29a
6.9 (1)
1
OH
2.3ꢃ0.2 (3)
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
OH
H
OH
H
H
H
289ꢃ2 (2)
790ꢃ10 (2)
2760ꢃ20 (2)
4295ꢃ295 (2)
242 (1)
960 (1)
2700 (1)
8500 (1)
9600 (1)
590 (1)
>1000 (1)
>1000 (1)
NDb
H
OH
H
H
OH
OH
H
H
Me
Me
Me
Ac
Ac
TBS
H
OH
H
>1000 (1)
>1000 (1)
91ꢃ27 (2)
289ꢃ17 (2)
>1000 (2)
37ꢃ4 (3)
OH
H
OH
H
OAc
H
NDb
OH
H
NDb
p-NO2(C6H4)CO2
OAc
H
NDb
H
H
23ꢃ2 (3)
NDb
MeOCO2
Me2NCO2
OMe
H
H
422ꢃ41 (2)
601 (1)
NDb
H
H
NDb
H
H
>1000 (1)
>1000 (2)
>1000 (2)
176ꢃ15 (2)
>1000 (2)
NDb
R1=R2=O
H
NDb
R1=R2=N(OMe)
H
H
NDb
OH
H
H
NDb
p-NO2(C6H4)CO2
H
NDb
a Cell growth inhibition under continuous exposure for 3–4 days, IC50ꢃrange nM (n) for n=2 or IC50ꢃSEM nM (n) for n=3.
b Not determined.

578
B. M. Gallagher, Jr. et al. / Bioorg. Med. Chem. Lett. 14 (2004) 575–579
analogous manner to 27–34 starting from the C.15 epi
diastereomer of 16 (not shown) utilizing (ꢀ)-DIPT for
the epoxidation. The C.15 stereochemistry of 36 was set
under standard Mitsunobu conditions (Ph3P, DEAD) in
the presence of the epoxide.
all appear to be important for activity. Also, there
appears to be some flexibility in the stereochemistry at
C.15, but an alcohol or isostere at this position may be
important. Derivatizations of the C.21–C.27 side–chain
and C.5–C.9 dihydropyran were not explored in the
current strategy, but may lead to analogues with
increased potencies. The results of these efforts will be
reported in due course.
All final compounds were evaluated in the MDA-MB-
435 human breast cancer cell line, and in some cases the
HT-29 human colon cancer cell line, for growth inhibi-
tory activities under continuous exposure conditions.9
The results are summarized in Table 1.
Acknowledgements
The tendency for the epoxide to be opened by the C.20
alcohol1b,c presents a potential liability for drug devel-
opment. To address this, a series of analogues were
prepared to explore the possibility of eliminating this
isomerization. Replacement of the epoxide of 1 with an
alkene (18) resulted in a loss in potency of two orders of
magnitude, indicating that this functionality is either
mechanistically or conformationally important for activ-
ity. Capping the C.20 alcohol with a methyl group (22)
resulted in a similar loss of potency. Interestingly, the
C.20 acetoxy analogue (25) was more potent than 22,
and the C.20 TBS ether (27) was inactive up to 1 mM.
This may indicate the importance of the C.20 alcohol to
participate in H-bonding interactions or that steric bulk
at this position is not tolerated. The reduction in
potencies of C.20 derivatives relative to 1 may also be
due to conformational changes in the side-chain and/or
the macrocycle. Regardless, these analogues indicate
that the C.20 alcohol plays an important role in biolo-
gical activity.
We thank Ms. Erin Murphy and Ms. Jiayi Wu for
technical assistance and Dr. Lynn D. Hawkins for cri-
tical reading of this manuscript.
References and notes
1. (a) Quinoa, E.; Kakou, Y.; Crews, P. J. Org. Chem. 1988,
53, 3642. (b) Corley, D. G.; Herb, R.; Moore, R. E.;
Scheuer, P. J.; Paul, V. J. J. Org. Chem. 1988, 53, 3644. (c)
Tanaka, J.-I.; Higa, T.; Bernardinelli, G.; Jefford, C. W.
Chem. Lett. 1996, 255. (d) Jefford, C. W.; Bernardinelli,
G.; Tanaka, J.-I.; Higa, T. Tetrahedron Lett. 1996, 37,
159.
2. (a) Mooberry, S. L.; Tien, G.; Hernandez, A. H.; Plu-
brukarn, A.; Davidson, B. S. Cancer Res. 1999, 59, 653.
(b) Mooberry, S. L.; Davidson, B. S. WO 0154689, 2001.
3. Pryor, D. E.; O’Brate, A.; Bilcer, G.; Diaz, J. F.; Wang,
Y.; Wang, Y.; Kabaki, M.; Jung, M. K.; Andreu, J. M.;
Ghosh, A. K.; Giannakakou, P.; Hamel, E. Biochemistry
2002, 41, 9109.
4. (a) Ghosh, A. K.; Wang, Y. J. Am. Chem. Soc. 2000, 122,
11027. (b) Ghosh, A. K.; Wang, Y. Tetrahedron Lett.
2001, 42, 3399. (c) Ghosh, A. K.; Wang, Y.; Kim, J. T.
J. Org. Chem. 2001, 66, 8973. (d) Paterson, I.; De Savi, C.;
Tudge, M. Org. Lett. 2001, 3, 3149. (e) Mulzer, J.; Ohler,
E. Angew. Chem., Int. Ed. 2001, 40, 3842. (f) Enev, V. E.;
Kahlig, H.; Mulzer, J. J. Am. Chem. Soc. 2001, 123,
10764. (g) Mulzer, J.; Hanbauer, M. Tetrahedron Lett.
2002, 43, 3381. (h) Wender, P. A.; Hegde, S. G.; Hubbard,
R. D.; Zhang, L. J. Am. Chem. Soc. 2002, 124, 4956. (i)
Crimmins, M. T.; Stanton, M. G.; Allwein, S. P. J. Am.
Chem. Soc. 2002, 124, 5958. (j) Williams, D. R.; Mi, L.;
Mullins, R. J.; Stites, R. E. Tetrahedron Lett. 2002, 43,
4841. (k) Nelson, S. G.; Cheung, W. S.; Kassick, A. J.;
Hilfiker, M. A. J. Am. Chem. Soc. 2002, 124, 13654. See
also Mulzer, J.; Enev, V. S. EP 1295886, 2003; Ashley, G.;
Metcalf, B. WO 02064589, 2002; Ghosh, A. K. WO
03076445, 2003. For a recent review, see: Crimmins, M. T.
Curr. Opin. Drug Disc. Dev. 2002, 5, 944.
5. (a) Shimizu, A.; Nishiyama, S. Tetrahedron Lett. 1997, 38,
6011. (b) Ghosh, A. K.; Mathivanan, P.; Cappiello, J.
Tetrahedron Lett. 1997, 38, 2427. (c) Shimizu, A.; Nish-
iyama, S. Synlett 1998, 1209. (d) Mulzer, J.; Hanbauer,
M. Tetrahedron Lett. 2000, 41, 33. (e) Ghosh, A. K.;
Wang, Y. Tetrahedron Lett. 2000, 41, 2319. (f) Ghosh,
A. K.; Wang, Y. Tetrahedron Lett. 2000, 41, 4705. (g)
Dorling, E. K.; Ohler, E.; Mulzer, J. Tetrahedron Lett.
2000, 41, 6323. (h) Paterson, I.; De Savi, C.; Tudge, M.
Org. Lett. 2001, 3, 213. (i) Nadolski, G. T.; Davidson,
B. S. Tetrahedron Lett. 2001, 42, 797. (j) Messenger, B. T.;
Davidson, B. S. Tetrahedron Lett. 2001, 42, 801. (k) Dor-
ling, E. K.; Ohler, E.; Mantoulidis, A.; Mulzer, J. Synlett
2001, 1105. (l) Ahmed, A.; Ohler, E.; Mulzer, J. Synthesis
2001, 2007. (m) Lee, H. W.; Jeong, C.-S.; Yoon, S. H.;
Inversion of the alcohol stereochemistry at C.15 resulted
in only a minor reduction in potency (cf. 18 versus 19
and 20 versus 21) indicating some tolerance for modifi-
cation at this position. To explore this further a series of
C.15 alcohol derivatives were prepared (28–32), but all
the analogues exhibit a loss in potency relative to 1.
Replacement of the C.15–C.17 part of the molecule with
a Michael accepting substructure (cf. 33 and 34) resulted
in inactive compounds up to 1 mM. Interestingly, the
C.15–C.17 tri-epi analogue (35) still remained fairly
potent. Thus while the absolute stereochemistry of C.15
may be of minor importance, the alcohol at this position
appears to contribute to the potency.
The C.2–C.3 Z enoate also plays an important role in
potency. The C.2–C.3 E enoate results in an approximate
5- to 10-fold loss in potency (cf. 20 versus 18 and 21
versus 19), and the C.2–C.3 alkynoate 23 or saturated
compound 24 are inactive. Whether this part of the
molecule is part of the pharmacophore or simply neces-
sary for macrocycle conformation is unclear at present.
In summary, a route to laulimalide was identified, which
enabled the synthesis of significant quantities of 1 and
related analogues to begin to explore the SAR. The key
steps include an indium mediated coupling to form the
C.14–C.15 bond, and a Yamaguchi macrolactonization
or H-W-E ring closure. All the analogues prepared
exhibited decreased potencies relative to 1. The C.16–
C.17 epoxide, the C.20 alcohol, and the C.2–C.3 enoate

B. M. Gallagher, Jr. et al. / Bioorg. Med. Chem. Lett. 14 (2004) 575–579
579
Lee, I. C. Bull. Korean Chem. Soc. 2001, 22, 791. (n) Lee,
H. W.; Yoon, S. H.; Lee, I. C.; Chung, B. Y. Bull. Korean
Chem. Soc. 2001, 22, 1179. (o) Sivaramakrishnan, A.;
Nadolski, G. T.; McAlexander, I. A.; Davidson, B. S.
Tetrahedron Lett. 2002, 43, 213.
7. All yields are averages from multiple preparative-scale
experiments.
8. Motoyoshi, H.; Ishigami, K.; Kitahara, T. Tetrahedron
2001, 57, 3899.
9. Towle, M. J.; Salvato, K. A.; Budrow, J.; Wels, B. F.;
Kuznetsov, G.; Aalfs, K. K.; Welsh, S.; Zheng, W.;
Seletsky, B. M.; Palme, M. H.; Habgood, G. J.; Singer,
L. A.; DiPietro, L. V.; Wang, Y.; Chen, J. J.; Quincy,
D. A.; Davis, A.; Yoshimatsu, K.; Kishi, Y.; Yu, M. J.;
Littlefield, B. A. Cancer Res. 2001, 61, 1013.
6. (a) Ahmed, A.; Hoegenauer, E. K.; Enev, V. S.; Han-
bauer, M.; Kaehlig, H.; Ohler, E.; Mulzer, J. J. Org.
Chem. 2003, 68, 3026. (b) Wender, P. A.; Hegde, S. G.;
Hubbard, R. D.; Zhang, L.; Mooberry, S. L. Org. Lett.
2003, 5, 3507.